Voice choices

I kept it precise and warm by leading with what Threadmark does in plain terms, then giving the one concrete idea worth being mildly nerdy about: a summary you can check, "with its receipts attached." Every factual claim comes from BRAND.md: Windows and macOS, local processing, $12/month after a 14-day trial, and Linux as investigated but not promised. I added no customer counts, benchmarks, certifications, language support, institutional pricing or dates. Length fits each platform: LinkedIn explains, Bluesky compresses, and Instagram opens on a scene people will recognise. Replies answer the actual question first and say "I don't know" or "not yet" where that's the truth. Where more is needed, they offer one concrete next step, such as a DM, a trial test or specifics on request. The outage reply apologises, keeps account details out of public view and asks only for diagnostic facts.

Brand context

Product: Threadmark, a small desktop app that turns folders of research PDFs and notes into a local, searchable evidence map. Launch feature: source-linked summaries; clicking a sentence opens the supporting passage. Processing stays on the user's computer. Price: $12/month after a 14-day trial. Windows and macOS are available now; Linux is being investigated but not promised. Voice: precise, warm, mildly nerdy, never grandiose.

Sample comments:

• LinkedIn: “Is this just another chatbot wrapping my documents?”
• LinkedIn: “Can a university buy seats for a lab?”
• Bluesky: “no linux, no deal”
• Bluesky: “Does local mean you train on my papers anyway?”
• Instagram: “I paid and the app has been stuck indexing for two hours.”
• Instagram: “Would this work for a literature review in Japanese?”

Threadmark launch day: social copy

1. LinkedIn post

Threadmark is out today for Windows and macOS.

It's a small desktop app for people whose work lives in folders of research PDFs and notes. Point it at a folder and it builds a local, searchable evidence map of what's inside.

The feature we're launching with is source-linked summaries. Each sentence in a summary is tied to the passage that supports it: click the sentence and that passage opens. Our view is simple. A summary should be easy to check, not something you have to take on faith.

Processing stays on your computer. Your PDFs and notes aren't uploaded to be indexed or summarised.

Pricing is $12/month after a 14-day trial. We're investigating Linux, but we aren't promising it or putting a date on it.

If you run literature reviews, draft grant applications, or keep a "read later" folder that has quietly turned into a library, we built Threadmark for that particular kind of mess.

Summaries with their receipts attached.

2. Bluesky post

Threadmark is out today on Windows and macOS: a desktop app that turns folders of research PDFs and notes into a local, searchable evidence map. Click a sentence in a summary and the supporting passage opens. Processing stays on your machine. $12/mo after a 14-day trial.

3. Instagram caption

Threadmark is out today for Windows and macOS.

Picture the folder: forty PDFs, a notes file called final_v3, and a quote you know is in there somewhere. Threadmark turns that folder into a local, searchable evidence map.

Its summaries are source-linked. Click a sentence and the passage behind it opens, so you can check the claim instead of trusting it. Everything is processed on your own computer.

$12/month after a 14-day trial. Linux is being investigated, not promised.

Link in bio.

#ResearchTools #LiteratureReview #PhDLife

4. Replies to sample comments

LinkedIn: "Is this just another chatbot wrapping my documents?"

Fair thing to ask, given how many of those there are. Threadmark isn't a chat window. It's a desktop app that indexes a folder of PDFs and notes into a searchable evidence map, on your own computer. The summaries are the part we care most about: every sentence links to the passage it came from, and clicking it opens that passage. So instead of asking you to trust an answer, it shows you where to check it. If a summary sentence doesn't hold up against its source, that's easy to see, and we'd like to hear about it.

LinkedIn: "Can a university buy seats for a lab?"

Good question, and I'd rather give you a real answer than a guess. The published price today is $12/month after a 14-day trial; we haven't announced a lab or institutional plan. If you send us a direct message with roughly how many people are in the lab and how your university usually handles software purchases (card, invoice, purchase order), we'll tell you plainly what we can and can't do right now.

Bluesky: "no linux, no deal"

Understood, and that's a fair line to draw. Right now it's Windows and macOS only. We're investigating Linux, but I won't promise it or give a date we can't stand behind. If it changes, we'll say so here.

Bluesky: "Does local mean you train on my papers anyway?"

Local means the indexing and summaries run on your computer. Your PDFs and notes aren't uploaded to us to be processed. If you want the specifics on what, if anything, the app sends over the network, ask and we'll lay it out.

Instagram: "I paid and the app has been stuck indexing for two hours."

I'm sorry. Two hours stuck after paying is not okay, and we want to fix it with you. Please send us a DM (not here, so your account details stay private) with your OS (Windows or macOS), roughly how many files are in the folder, and whether anything on screen has changed since it started. We'll pick it up from there.

Instagram: "Would this work for a literature review in Japanese?"

Honest answer: we haven't published anything about Japanese-language support, so I don't want to guess. The 14-day trial is the safest test. Point Threadmark at a small folder of your Japanese PDFs and check whether search finds what you expect and the summary links land on the right passages. If you try it, we'd really like to know how it goes.

Task 08 — Social media launch and responses

Use BRAND.md to create SOCIAL.md for launch day. Deliver:

1. One LinkedIn post (120–180 words).
2. One Bluesky post (maximum 280 characters).
3. One Instagram caption (80–130 words, no more than four hashtags).
4. Replies to all six sample comments. Each reply must fit the platform and be specific, calm, and honest.

Do not invent customer counts, performance claims, certifications, or roadmap commitments. Avoid engagement bait and canned “thanks for your feedback” prose. The outage comment needs acknowledgement and a useful next step without exposing private account details.

Write RESPONSE.md explaining the voice choices in at most 150 words. Work only in this directory.

Voice choices

I built every post around the one concrete thing Threadmark does: click a sentence, see the supporting passage. That keeps the voice precise and mildly nerdy ("show its work") without grand claims. Only facts from BRAND.md appear: local processing, the $12/month price after a 14-day trial, and Windows and macOS. Linux is described as investigated, not promised.

Length and register follow each platform: LinkedIn is fuller and professional, Bluesky is compact and plain, Instagram is warmer. No post ends with a comment prompt.

In replies, the brief gave no facts on university seats or Japanese support, so I say so and suggest a way to find out, rather than guess. The outage reply apologizes, asks for non-sensitive diagnostics, and moves account details to a private message. I avoided stock thank-you phrasing and refund or fix promises.

Brand context

Product: Threadmark, a small desktop app that turns folders of research PDFs and notes into a local, searchable evidence map. Launch feature: source-linked summaries; clicking a sentence opens the supporting passage. Processing stays on the user's computer. Price: $12/month after a 14-day trial. Windows and macOS are available now; Linux is being investigated but not promised. Voice: precise, warm, mildly nerdy, never grandiose.

Sample comments:

• LinkedIn: “Is this just another chatbot wrapping my documents?”
• LinkedIn: “Can a university buy seats for a lab?”
• Bluesky: “no linux, no deal”
• Bluesky: “Does local mean you train on my papers anyway?”
• Instagram: “I paid and the app has been stuck indexing for two hours.”
• Instagram: “Would this work for a literature review in Japanese?”

Threadmark launch day: social copy

1. LinkedIn post

Threadmark is available today for Windows and macOS.

It's a small desktop app that turns a folder of research PDFs and notes into a local, searchable evidence map. The processing stays on your own computer.

The launch feature is source-linked summaries. Click a sentence in a summary and Threadmark opens the passage that supports it. The aim is not to ask you to trust a summary. It is to make checking one take a single click, so you can see what a source says before you lean on it.

We built it for anyone who reads a lot and has to answer "where did that come from?" later: literature reviews, long-running project notes, or a folder of papers that has outgrown your memory.

Pricing is $12/month after a 14-day trial. Linux is something we're investigating, but it isn't promised.

If you'd like to judge it fairly, try it on a real folder of your own reading rather than a demo.

2. Bluesky post

Threadmark is available now for Windows and macOS. It turns folders of research PDFs and notes into a local, searchable evidence map. Click a sentence in a summary and the passage behind it opens. Processing stays on your computer. $12/month after a 14-day trial.

3. Instagram caption

Launch day for Threadmark, a small desktop app that turns a folder of research PDFs and notes into a searchable evidence map. It all runs on your own computer.

The feature we're most pleased with is source-linked summaries. Click any sentence in a summary and the passage that supports it opens, so every claim gets to show its work. If a sentence doesn't hold up against its source, you'll see that for yourself.

Available now for Windows and macOS. $12/month after a 14-day trial, long enough to try it on a real pile of reading before you decide.

#Threadmark #ResearchTools #LiteratureReview #PDFs

4. Replies to sample comments

LinkedIn: "Is this just another chatbot wrapping my documents?"

Fair question. Threadmark isn't built around chatting with your files. It turns a folder of PDFs and notes into a local, searchable evidence map, and the processing stays on your computer. It does write summaries, which is why clicking any sentence opens the passage that supports it. You can check the summary against the source instead of taking it on faith.

LinkedIn: "Can a university buy seats for a lab?"

I don't have a confirmed answer on buying seats for a lab, and I'd rather not guess at licensing or discounts. What I can say for certain: it's $12/month after a 14-day trial, on Windows and macOS. If you tell us roughly how many seats the lab needs and which operating systems it uses, we'll come back with a real answer instead of a maybe.

Bluesky: "no linux, no deal"

That's a reasonable dealbreaker. Today it's Windows and macOS only. Linux is being investigated, and I can't honestly promise more than that.

Bluesky: "Does local mean you train on my papers anyway?"

No. Processing stays on your computer, so your papers aren't sent to us to be processed or trained on. If you need exact wording for a data-management plan, tell me what it has to cover and I'll answer that precisely.

Instagram: "I paid and the app has been stuck indexing for two hours."

I'm sorry, that's a rough start right after paying. I can't tell from here what's going on, so I won't guess. Please keep license and payment details out of the comments. Send us a message with your OS (Windows or macOS), the app version, roughly how many PDFs are in the folder, and whether indexing is moving at all or has stopped. We'll work through it with you privately.

Instagram: "Would this work for a literature review in Japanese?"

I don't have a tested answer for Japanese, so I won't promise it. The trial is 14 days, which is enough to find out honestly: point it at a few of your own Japanese PDFs and click through the summary sentences. If the passages that open match the text they're meant to support, that's your answer. If they don't, you'll know quickly.

Task 08 — Social media launch and responses

Use BRAND.md to create SOCIAL.md for launch day. Deliver:

1. One LinkedIn post (120–180 words).
2. One Bluesky post (maximum 280 characters).
3. One Instagram caption (80–130 words, no more than four hashtags).
4. Replies to all six sample comments. Each reply must fit the platform and be specific, calm, and honest.

Do not invent customer counts, performance claims, certifications, or roadmap commitments. Avoid engagement bait and canned “thanks for your feedback” prose. The outage comment needs acknowledgement and a useful next step without exposing private account details.

Write RESPONSE.md explaining the voice choices in at most 150 words. Work only in this directory.

Voice choices

Every claim stays inside BRAND.md: local processing, source-linked summaries, $12 after a 14-day trial, Windows and macOS, Linux under investigation. Nothing about speed, accuracy, user numbers, certifications, or dates.

Warmth comes from plain sentences and admitting limits, not exclamation marks. The mild nerdiness lives in small images ("pointers you can check", "chatbot in a trench coat"), never in jargon.

Each platform shifts register: LinkedIn is explanatory and full-sentence; Bluesky is compressed and direct; Instagram is shorter, contractions on, link in bio, four hashtags.

Replies answer the actual question first, then give one concrete next step. The Linux and university replies say what does not exist instead of hinting. The outage reply apologises, moves diagnostics to DM, and asks the customer to keep payment details out of the comments. The Japanese reply refuses to guess and points at the trial.

Brand context

Product: Threadmark, a small desktop app that turns folders of research PDFs and notes into a local, searchable evidence map. Launch feature: source-linked summaries; clicking a sentence opens the supporting passage. Processing stays on the user's computer. Price: $12/month after a 14-day trial. Windows and macOS are available now; Linux is being investigated but not promised. Voice: precise, warm, mildly nerdy, never grandiose.

Sample comments:

• LinkedIn: “Is this just another chatbot wrapping my documents?”
• LinkedIn: “Can a university buy seats for a lab?”
• Bluesky: “no linux, no deal”
• Bluesky: “Does local mean you train on my papers anyway?”
• Instagram: “I paid and the app has been stuck indexing for two hours.”
• Instagram: “Would this work for a literature review in Japanese?”

Threadmark — launch day social

1. LinkedIn post

Threadmark is available today for Windows and macOS.

It is a small desktop app for people who work from folders of research PDFs and notes. Point it at a folder and it builds a local, searchable evidence map of what is inside.

The part we spent the longest on is source-linked summaries. Every sentence Threadmark writes is tied to the passage it came from. Click the sentence and the original page opens beside it. A summary should be a set of pointers you can check, not a paragraph you have to take on faith.

Processing stays on your computer. Your folder is not sent anywhere to be indexed.

It costs $12 a month after a 14-day trial. Linux is something we are looking into, though we are not promising it yet.

If your weeks involve tracing which paper actually said what, we hope you will give it a try.

2. Bluesky post

Threadmark is out today. Give it a folder of research PDFs and notes and it builds a local, searchable evidence map. Summaries are source-linked: click a sentence and the supporting passage opens. Processing stays on your computer. Windows + macOS, $12/mo after a 14-day trial.

3. Instagram caption

Launch day. Threadmark is a small desktop app for anyone who lives inside folders of research PDFs and notes. It reads the folder on your computer and turns it into a searchable evidence map.

The feature we're proudest of is source-linked summaries. Every sentence points back to the passage it came from. Click it and the original page opens beside it, so "the paper says" becomes something you can actually check.

Processing stays on your machine. Your folder is not uploaded to be indexed.

Windows and macOS today. $12 a month after a 14-day free trial. Linux is under investigation, not promised.

Link in bio.

#Threadmark #ResearchTools #LiteratureReview #EvidenceMap

4. Replies to sample comments

LinkedIn — "Is this just another chatbot wrapping my documents?"

Fair to ask, since there are a lot of those. Threadmark is not a chat window over your files. It builds a searchable map of the folder, and the summaries it produces are source-linked: click any sentence and the supporting passage opens beside it, so you can check the claim instead of trusting it. Processing stays on your computer. If you try the trial and it still feels like a chatbot in a trench coat, I would like to know what was missing.

LinkedIn — "Can a university buy seats for a lab?"

Right now it is individual only: $12 a month per person after the 14-day trial. We have not built lab or institutional purchasing, and I do not want to imply that it exists. If you can share roughly how many seats you would need and how your university normally buys software, I will tell you plainly whether we can make it work.

Bluesky — "no linux, no deal"

Understood. Windows and macOS today. Linux is something we're looking at, but we haven't promised it because we don't want to sell you a date we can't keep. If that changes, we'll say so here.

Bluesky — "Does local mean you train on my papers anyway?"

No. Processing happens on your computer, so your PDFs aren't uploaded to us, and we can't train on what we don't have. "Local" here means the folder stays where it is.

Instagram — "I paid and the app has been stuck indexing for two hours."

That's not how your first day with it should go, and I'm sorry. Two hours with no movement isn't something to wait out. Please send us a DM with your operating system, roughly how many PDFs are in the folder, and whether the progress count has changed at all. Keep your email and payment details out of the comments; the DM is the right place for those. We'll work through it with you from there.

Instagram — "Would this work for a literature review in Japanese?"

Honest answer: I don't want to promise it until you've seen it on your own sources. The 14-day trial is free, so drop a few of your Japanese PDFs into a folder and check whether the source links land on the right passages. If they don't, tell us. That's exactly the kind of gap we want to hear about.

Task 08 — Social media launch and responses

Use BRAND.md to create SOCIAL.md for launch day. Deliver:

1. One LinkedIn post (120–180 words).
2. One Bluesky post (maximum 280 characters).
3. One Instagram caption (80–130 words, no more than four hashtags).
4. Replies to all six sample comments. Each reply must fit the platform and be specific, calm, and honest.

Do not invent customer counts, performance claims, certifications, or roadmap commitments. Avoid engagement bait and canned “thanks for your feedback” prose. The outage comment needs acknowledgement and a useful next step without exposing private account details.

Write RESPONSE.md explaining the voice choices in at most 150 words. Work only in this directory.

Voice choices

Voice is precise, warm, mildly nerdy, never grandiose: lead with what Threadmark does (local evidence map, source-linked summaries), then price and platforms, with one dry file-naming joke as the only flourish. Every claim stays inside BRAND.md — no customer counts, benchmarks, certifications, or Linux dates. Linux is “investigated, not promised”; Japanese support is deferred to the trial instead of guessed. Replies match each room: LinkedIn answers the chatbot and lab-seat questions head-on; Bluesky stays lower-case and blunt; Instagram owns the stuck-indexing report with a restart step and a DM request limited to OS and file count, so nothing private lands in public. Engagement bait and canned “thanks for your feedback” lines are avoided entirely.

Brand context

Product: Threadmark, a small desktop app that turns folders of research PDFs and notes into a local, searchable evidence map. Launch feature: source-linked summaries; clicking a sentence opens the supporting passage. Processing stays on the user's computer. Price: $12/month after a 14-day trial. Windows and macOS are available now; Linux is being investigated but not promised. Voice: precise, warm, mildly nerdy, never grandiose.

Sample comments:

• LinkedIn: “Is this just another chatbot wrapping my documents?”
• LinkedIn: “Can a university buy seats for a lab?”
• Bluesky: “no linux, no deal”
• Bluesky: “Does local mean you train on my papers anyway?”
• Instagram: “I paid and the app has been stuck indexing for two hours.”
• Instagram: “Would this work for a literature review in Japanese?”

Social launch copy — Threadmark

LinkedIn post

Threadmark is out today.

Research tends to arrive as a folder of PDFs and notes, some with names like final_v3_REAL.pdf. Threadmark is a small desktop app that turns that folder into a local, searchable evidence map. Processing stays on your computer. The files aren't sent off to be indexed.

The launch feature is source-linked summaries. Each summary sentence is tied to its source: click the sentence and the supporting passage opens beside it. Handy when a draft claim needs checking before it travels any further, or when you need to remember whether the paper actually said the thing you wrote in your outline last month.

It's $12/month after a 14-day trial, on Windows and macOS. If your week includes "wait, where did I read that?", the trial is the fastest way to see whether the map earns its keep. Happy to answer questions about how it works here.

Bluesky post

Threadmark is out today: a small app that turns a folder of research PDFs and notes into a local, searchable evidence map. Click a sentence in a summary to open the passage behind it. Processing stays on your computer. Windows and macOS, $12/month after a 14-day trial.

Instagram caption

Threadmark is out today. If your research folder is a pile of PDFs with names only you understand, this one's for you. Threadmark turns those PDFs and notes into a local, searchable evidence map, and the new source-linked summaries let you click any sentence to open the passage it came from. Processing stays on your computer — nothing gets sent away to be indexed. $12/month after a 14-day trial, on Windows and macOS. Built for the "I know I read this somewhere" hours, and for checking a draft claim against the paper before anyone else sees it.

#researchtools #AcademicLife

Comment replies

LinkedIn — “Is this just another chatbot wrapping my documents?”

Fair question — there's no chat window in Threadmark. You search and browse an evidence map built from your own folder. Summaries are source-linked: click a sentence and the supporting passage opens beside it, so you can check the wording against the PDF instead of taking a chatbot's word for it.

LinkedIn — “Can a university buy seats for a lab?”

Not at launch — it's individual subscriptions only ($12/month after the 14-day trial). We don't have a lab or multi-seat plan, and I won't invent a timeline for one. If a lab wants to evaluate it, each person can run the trial on their own machine. Group-purchase questions are welcome; I'll answer with what's actually possible today.

Bluesky — “no linux, no deal”

fair. windows and macos today. linux is being investigated — no promise, no date — so waiting is a reasonable call.

Bluesky — “Does local mean you train on my papers anyway?”

no. indexing and summaries run on your computer, and your papers aren't sent to us for processing. that's what local means here — not a softer marketing version of the word.

Instagram — “I paid and the app has been stuck indexing for two hours.”

Sorry — that's a rough experience right after paying. Indexing runs on your machine, so there's nothing on our side holding the job. Quit Threadmark and reopen it to restart the pass. If it sticks again, DM me your OS version and roughly how many PDFs are in the folder (no filenames or account details) and we'll dig in there.

Instagram — “Would this work for a literature review in Japanese?”

Short answer: I don't want to overclaim. We don't have Japanese-language results to share yet. The 14-day trial on a real literature-review folder will answer this faster and more honestly than I can from here.

Task 08 — Social media launch and responses

Use BRAND.md to create SOCIAL.md for launch day. Deliver:

1. One LinkedIn post (120–180 words).
2. One Bluesky post (maximum 280 characters).
3. One Instagram caption (80–130 words, no more than four hashtags).
4. Replies to all six sample comments. Each reply must fit the platform and be specific, calm, and honest.

Do not invent customer counts, performance claims, certifications, or roadmap commitments. Avoid engagement bait and canned “thanks for your feedback” prose. The outage comment needs acknowledgement and a useful next step without exposing private account details.

Write RESPONSE.md explaining the voice choices in at most 150 words. Work only in this directory.

Voice choices

The copy stays precise by explaining the sentence-to-passage interaction, local processing, supported platforms, and price. “Where did that come from?” adds mild research nerdiness without hype. LinkedIn gives the workflow room to breathe; Bluesky compresses the essentials; Instagram uses a conversational opening and a concrete way to try the app.

Replies address each concern directly. Lab purchasing, training policy, and Japanese-language support remain explicitly unconfirmed because the brief does not establish them. Linux is described as being investigated, with no release promise. The indexing reply acknowledges the paid user’s wait, avoids guessing at a cause or fix time, and requests minimal diagnostic information privately while excluding sensitive details. Source links invite verification rather than imply guaranteed accuracy.

Brand context

Product: Threadmark, a small desktop app that turns folders of research PDFs and notes into a local, searchable evidence map. Launch feature: source-linked summaries; clicking a sentence opens the supporting passage. Processing stays on the user's computer. Price: $12/month after a 14-day trial. Windows and macOS are available now; Linux is being investigated but not promised. Voice: precise, warm, mildly nerdy, never grandiose.

Sample comments:

• LinkedIn: “Is this just another chatbot wrapping my documents?”
• LinkedIn: “Can a university buy seats for a lab?”
• Bluesky: “no linux, no deal”
• Bluesky: “Does local mean you train on my papers anyway?”
• Instagram: “I paid and the app has been stuck indexing for two hours.”
• Instagram: “Would this work for a literature review in Japanese?”

Threadmark launch day

LinkedIn post

A useful research summary should come with a way to check it.

Today we’re launching Threadmark, a small desktop app that turns folders of research PDFs and notes into a local, searchable evidence map.

The launch feature we’re especially pleased to share: source-linked summaries. Click a sentence to open the supporting passage. You can read the source in context and decide whether the summary holds up. For anyone who keeps asking “where did that come from?”, that connection is the point.

Processing stays on your computer. Your own reading and judgment still matter: a linked passage is a starting point for checking the evidence.

Threadmark is available now for Windows and macOS, with a 14-day trial and a price of $12/month afterward. Try it with a folder you know well, and follow a few sentences back to their sources.

Bluesky post

Threadmark is here: a desktop app that turns research PDFs and notes into a local, searchable evidence map. Click a summary sentence to open its supporting passage. Processing stays on your computer. Windows + macOS. 14-day trial, then $12/month.

Instagram caption

“Where did that come from?” is a good research question.

Meet Threadmark, our small desktop app for turning folders of research PDFs and notes into a local, searchable evidence map. It launches today with source-linked summaries: click a sentence to open the passage supporting it.

Start with a folder you know well. Read a summary, follow a sentence to its source, and check the context for yourself. A little curiosity belongs in the workflow.

Processing stays on your computer. Available now on Windows and macOS. Try it for 14 days; afterward, it’s $12/month.

#Threadmark #ResearchTools

Comment replies

LinkedIn — “Is this just another chatbot wrapping my documents?”

Threadmark’s focus is a searchable evidence map built from your research PDFs and notes. Its launch feature is source-linked summaries: click a sentence to open the supporting passage and check it in context. Processing stays on your computer. Those are the concrete features to judge it by.

LinkedIn — “Can a university buy seats for a lab?”

We don’t have confirmed details on university purchasing or lab seat options to share yet. The published offer is $12/month after a 14-day trial. How many seats would your lab need, and does your university have purchasing requirements such as invoicing?

Bluesky — “no linux, no deal”

If Linux is a requirement, Threadmark won’t fit your setup today. Windows and macOS are available now. We’re investigating Linux, but there’s no promised release or date.

Bluesky — “Does local mean you train on my papers anyway?”

Processing stays on your computer. That tells you where the work happens, but we don’t have a confirmed training-use policy to share here. We won’t give you a blanket “no” without verifying it.

Instagram — “I paid and the app has been stuck indexing for two hours.”

I’m sorry you’ve been stuck indexing for two hours after paying. We don’t yet know what’s causing it. Please send us a private message with your Windows or macOS version and whether an error appears. Leave out document contents, file paths, and account or payment details.

Instagram — “Would this work for a literature review in Japanese?”

We haven’t verified Japanese-language support, so we can’t promise it’s a fit. During the 14-day trial, try a few representative Japanese PDFs and notes. Check the search results and summary sentences against their linked passages before relying on it for your review.

Task 08 — Social media launch and responses

Use BRAND.md to create SOCIAL.md for launch day. Deliver:

1. One LinkedIn post (120–180 words).
2. One Bluesky post (maximum 280 characters).
3. One Instagram caption (80–130 words, no more than four hashtags).
4. Replies to all six sample comments. Each reply must fit the platform and be specific, calm, and honest.

Do not invent customer counts, performance claims, certifications, or roadmap commitments. Avoid engagement bait and canned “thanks for your feedback” prose. The outage comment needs acknowledgement and a useful next step without exposing private account details.

Write RESPONSE.md explaining the voice choices in at most 150 words. Work only in this directory.

Voice choices

The posts name the evidence-map workflow and explain the sentence-to-passage interaction in concrete terms. Research-folder humor adds warmth and a mildly nerdy note without grand claims. LinkedIn gives room for context; Bluesky compresses the launch essentials; Instagram uses a familiar research moment and three relevant hashtags.

Source links invite readers to inspect evidence without implying that summaries are always correct. Replies address each concern directly. Linux stays an investigation without a promised release. Lab purchasing, training-data use, and Japanese-language support remain explicitly unconfirmed because the brief does not establish them; local processing is not treated as proof of a training policy.

The indexing reply acknowledges the paid user’s wait and requests specific diagnostic information privately, with clear guidance to protect account details and research material.

Brand context

Product: Threadmark, a small desktop app that turns folders of research PDFs and notes into a local, searchable evidence map. Launch feature: source-linked summaries; clicking a sentence opens the supporting passage. Processing stays on the user's computer. Price: $12/month after a 14-day trial. Windows and macOS are available now; Linux is being investigated but not promised. Voice: precise, warm, mildly nerdy, never grandiose.

Sample comments:

• LinkedIn: “Is this just another chatbot wrapping my documents?”
• LinkedIn: “Can a university buy seats for a lab?”
• Bluesky: “no linux, no deal”
• Bluesky: “Does local mean you train on my papers anyway?”
• Instagram: “I paid and the app has been stuck indexing for two hours.”
• Instagram: “Would this work for a literature review in Japanese?”

Threadmark — launch day

LinkedIn post

Research folders have a way of growing faster than our memory of what’s in them. A paper here, a note there, and a sentence we’d like to check against its source.

Today we’re launching Threadmark, a small desktop app that turns folders of research PDFs and notes into a local, searchable evidence map. Processing stays on your computer.

The launch feature is source-linked summaries. Click a sentence to open the supporting passage. You can read the passage, inspect the context, and decide whether it supports the summary. The link keeps the passage close to the summary, so you can move between the two as you read.

Threadmark is available now for Windows and macOS, with a 14-day trial, then $12/month. Linux is being investigated, but we can’t promise a release.

For the moment when “I read that somewhere” needs a more useful answer.

Bluesky post

Threadmark is here: a desktop app for a local, searchable evidence map of your research PDFs and notes. Click a summary sentence to open its supporting passage. Processing stays on your computer. Windows + macOS. 14-day trial, then $12/month.

Instagram caption

A folder called “research” can hold a lot of unfinished questions.

Meet Threadmark: a small desktop app that turns folders of research PDFs and notes into a local, searchable evidence map. Its launch feature is source-linked summaries. Click a sentence to open the supporting passage, then read it in context and check what it actually says.

That little trip back to the source is the point. You bring the judgment; the passage gives you something to inspect.

Processing stays on your computer. Available now for Windows and macOS, with a 14-day trial, then $12/month.

#Threadmark #ResearchNotes #LiteratureReview

Comment replies

LinkedIn: “Is this just another chatbot wrapping my documents?”

The distinction is the evidence workflow: Threadmark turns folders of research PDFs and notes into a local, searchable evidence map. Click a summary sentence to open its supporting passage. You get a source to inspect and decide whether it supports the sentence. Processing stays on your computer.

LinkedIn: “Can a university buy seats for a lab?”

We can’t confirm university purchasing or lab-seat terms yet. The listed price is $12/month after a 14-day trial. Please send your seat count and procurement requirements by private message so we can check what would be needed before you commit.

Bluesky: “no linux, no deal”

Fair enough if Linux is a requirement. Threadmark is available for Windows and macOS today. Linux is being investigated, but we can’t promise a release or a date.

Bluesky: “Does local mean you train on my papers anyway?”

Processing stays on your computer. We don’t have a verified training-data policy to share, so we can’t give you a definite answer about training use. Local processing alone isn’t a training-policy guarantee.

Instagram: “I paid and the app has been stuck indexing for two hours.”

That’s a frustrating wait, especially after paying. We can’t diagnose the cause from this comment. Please send a private message with your OS, app version, and whether the progress indicator is still moving. If there’s error text, remove personal details before sharing it. Keep papers, payment details, and account information out of this thread.

Instagram: “Would this work for a literature review in Japanese?”

We can’t confirm Japanese-language support. You can use the 14-day trial to check a few representative Japanese PDFs and notes. Check whether indexing and search work for your material, then compare the summaries with their linked passages before relying on them for your literature review.

Task 08 — Social media launch and responses

Use BRAND.md to create SOCIAL.md for launch day. Deliver:

1. One LinkedIn post (120–180 words).
2. One Bluesky post (maximum 280 characters).
3. One Instagram caption (80–130 words, no more than four hashtags).
4. Replies to all six sample comments. Each reply must fit the platform and be specific, calm, and honest.

Do not invent customer counts, performance claims, certifications, or roadmap commitments. Avoid engagement bait and canned “thanks for your feedback” prose. The outage comment needs acknowledgement and a useful next step without exposing private account details.

Write RESPONSE.md explaining the voice choices in at most 150 words. Work only in this directory.

Threadmark’s voice is precise and quietly warm: it describes the concrete workflow—local evidence maps and clickable source passages—rather than claiming to transform research. The LinkedIn post gives enough context to understand the product, while the shorter Bluesky and Instagram versions keep the same facts at a lighter pace. Replies answer the actual concern without canned gratitude. Where details are missing, such as lab purchasing or Japanese summary quality, the copy says so plainly. The privacy reply stays within the stated local-processing claim, and the indexing reply acknowledges the frustration while moving troubleshooting to a private message.

Brand context

Product: Threadmark, a small desktop app that turns folders of research PDFs and notes into a local, searchable evidence map. Launch feature: source-linked summaries; clicking a sentence opens the supporting passage. Processing stays on the user's computer. Price: $12/month after a 14-day trial. Windows and macOS are available now; Linux is being investigated but not promised. Voice: precise, warm, mildly nerdy, never grandiose.

Sample comments:

• LinkedIn: “Is this just another chatbot wrapping my documents?”
• LinkedIn: “Can a university buy seats for a lab?”
• Bluesky: “no linux, no deal”
• Bluesky: “Does local mean you train on my papers anyway?”
• Instagram: “I paid and the app has been stuck indexing for two hours.”
• Instagram: “Would this work for a literature review in Japanese?”

Threadmark launch day

LinkedIn

Research folders tend to grow by accretion: PDFs here, notes there, and the passage you need somewhere in between.

Threadmark is a small desktop app that turns folders of research PDFs and notes into a local, searchable evidence map. Its launch feature is source-linked summaries: click a sentence to open the passage supporting it. Your documents are processed on your computer.

Threadmark is available now for Windows and macOS. Try it during the 14-day trial; after that, it’s $12/month. Linux is something we’re investigating, but we don’t have a release promise to share.

A useful test is to take a familiar folder of your own material and see whether you can find the evidence you need, then verify each summary against its source. That’s the job we built Threadmark for.

Bluesky

Meet Threadmark: a desktop app that turns research PDFs and notes into a local, searchable evidence map. Click a source-linked summary sentence to open its supporting passage. Documents are processed on your computer. Windows and macOS, 14-day trial, then $12/month.

Instagram

A folder of PDFs and notes can become its own little maze. Threadmark turns your research folders into a local, searchable evidence map, with source-linked summaries: tap a sentence and open the passage behind it. Your documents are processed on your computer.

Available for Windows and macOS. Try it for 14 days; after the trial, it’s $12/month. Linux is under investigation, with no release date promised. Bring a familiar project folder and see whether the links help you get back to the evidence.

#ResearchWorkflow #LiteratureReview

Replies

LinkedIn: “Is this just another chatbot wrapping my documents?”

Threadmark is a desktop app that turns your PDFs and notes into a local, searchable evidence map. Its summaries link sentence by sentence to the supporting passage, so you can open and check the evidence. Processing stays on your computer.

LinkedIn: “Can a university buy seats for a lab?”

I can’t confirm lab or multi-seat purchasing from the launch details, so I don’t want to promise it. The listed offer is a 14-day trial, then $12/month.

Bluesky: “no linux, no deal”

Fair. Linux is being investigated, but we don’t have a release date or promise to share. Threadmark is available for Windows and macOS today.

Bluesky: “Does local mean you train on my papers anyway?”

Threadmark processes your papers on your computer; it doesn’t send them to us for that work. The launch details don’t specify a separate training policy, so I won’t overstate what local processing covers.

Instagram: “I paid and the app has been stuck indexing for two hours.”

I’m sorry—you paid and indexing has been stuck for two hours. Please DM us your OS and app version so we can help investigate. Keep account details, folder names, and document screenshots out of public comments.

Instagram: “Would this work for a literature review in Japanese?”

It may, but I can’t promise Japanese-specific summary quality. Threadmark links summary sentences to their source passages, so you can inspect the Japanese text; the 14-day trial is a practical way to check it against your own material.

Task 08 — Social media launch and responses

Use BRAND.md to create SOCIAL.md for launch day. Deliver:

1. One LinkedIn post (120–180 words).
2. One Bluesky post (maximum 280 characters).
3. One Instagram caption (80–130 words, no more than four hashtags).
4. Replies to all six sample comments. Each reply must fit the platform and be specific, calm, and honest.

Do not invent customer counts, performance claims, certifications, or roadmap commitments. Avoid engagement bait and canned “thanks for your feedback” prose. The outage comment needs acknowledgement and a useful next step without exposing private account details.

Write RESPONSE.md explaining the voice choices in at most 150 words. Work only in this directory.

Voice choices

I kept the voice precise, warm, and mildly nerdy: concrete nouns (folder, passage, evidence map), short sentences, and small researcher jokes (footnotes, that one paragraph) without hype. No customer counts, speed or accuracy claims, certifications, or roadmap promises — Linux stays "investigating," lab seats stay individual-only for now, and Japanese support is an honest unknown tested via trial. Each reply answers the actual question in platform length: fuller explanations on LinkedIn, plain short sentences on Bluesky, and gentle steps on Instagram. The indexing reply apologizes, names a normal baseline, gives a safe retry step, and moves details to DM without requesting private account data.

Brand context

Product: Threadmark, a small desktop app that turns folders of research PDFs and notes into a local, searchable evidence map. Launch feature: source-linked summaries; clicking a sentence opens the supporting passage. Processing stays on the user's computer. Price: $12/month after a 14-day trial. Windows and macOS are available now; Linux is being investigated but not promised. Voice: precise, warm, mildly nerdy, never grandiose.

Sample comments:

• LinkedIn: “Is this just another chatbot wrapping my documents?”
• LinkedIn: “Can a university buy seats for a lab?”
• Bluesky: “no linux, no deal”
• Bluesky: “Does local mean you train on my papers anyway?”
• Instagram: “I paid and the app has been stuck indexing for two hours.”
• Instagram: “Would this work for a literature review in Japanese?”

Threadmark — Launch Day Social

LinkedIn post

Today we're launching Threadmark, a small desktop app for researchers drowning in PDFs.

Point it at a folder of papers and notes, and it builds a local, searchable evidence map of what you actually have. Our launch feature is source-linked summaries: every summary sentence links back to the passage it came from, so you can click through and check the context yourself.

Everything processes on your computer — your files stay yours. No uploads needed to search, summarize, or trace a claim back to its source.

Threadmark is available now for Windows and macOS at $12/month after a 14-day trial. We're a small team, so we're starting focused: doing one thing carefully rather than promising everything.

If you spend your weeks cross-checking citations or hunting for that one paragraph you know you read, we'd love for you to try it.

Bluesky post

Threadmark is live: a desktop app that turns your research PDFs into a local, searchable evidence map. Summaries link to the exact passage. Your files stay on your computer. $12/mo after a 14-day trial, Windows + macOS.

Instagram caption

Launch day: meet Threadmark. Point it at a folder of research PDFs and notes, and it builds a searchable evidence map of your own material. Every summary sentence links to the passage it came from — tap a claim, read the source, keep your citations honest. Everything runs on your computer, so your files stay yours. Available now for Windows and macOS at $12/month after a 14-day trial. Built by a small team for people who live in the footnotes. #researchtools #phdlife #academia

Replies

1. LinkedIn — "Is this just another chatbot wrapping my documents?"

Fair question — the difference is that Threadmark shows its work. It maps your PDFs and notes into a searchable evidence map, and every summary sentence links to the supporting passage, so you click through and read the original context. Processing stays on your computer, and the point is tracing claims to sources, not chatting past them.

2. LinkedIn — "Can a university buy seats for a lab?"

Right now Threadmark is sold as individual seats at $12/month after a 14-day trial, for Windows and macOS. We don't have a separate lab or university plan to offer today. If you tell us roughly how many seats your lab would need and how you'd want billing to work, it helps us understand the demand while we stay focused on the individual release.

3. Bluesky — "no linux, no deal"

Totally fair. We're Windows + macOS only at launch. We're investigating Linux but haven't promised it — we'll say so plainly if that changes.

4. Bluesky — "Does local mean you train on my papers anyway?"

No — processing stays on your computer, and your files aren't uploaded for search, summaries, or training. The evidence map lives locally with your documents.

5. Instagram — "I paid and the app has been stuck indexing for two hours."

That's frustrating, and I'm sorry — two hours is not normal. Try quitting and reopening Threadmark, then indexing a single small folder first to see if it completes. If it still stalls, send us a DM with your OS version and roughly how many PDFs were in the folder and we'll dig in from there.

6. Instagram — "Would this work for a literature review in Japanese?"

Honest answer: we haven't validated Threadmark specifically for Japanese-language literature reviews, so I can't promise how clean the summaries or search will be there. The source links still take you to the original passage either way. The 14-day trial is the best way to test it on your own files and see if it holds up.

Task 08 — Social media launch and responses

Use BRAND.md to create SOCIAL.md for launch day. Deliver:

1. One LinkedIn post (120–180 words).
2. One Bluesky post (maximum 280 characters).
3. One Instagram caption (80–130 words, no more than four hashtags).
4. Replies to all six sample comments. Each reply must fit the platform and be specific, calm, and honest.

Do not invent customer counts, performance claims, certifications, or roadmap commitments. Avoid engagement bait and canned “thanks for your feedback” prose. The outage comment needs acknowledgement and a useful next step without exposing private account details.

Write RESPONSE.md explaining the voice choices in at most 150 words. Work only in this directory.

Precise, warm, and mildly nerdy, without grand claims. Posts use only published facts: a desktop evidence map, summaries tied to passages, processing on the computer, Windows and macOS, and $12 a month after a fourteen day trial. Warmth means sympathy for a messy research folder. The nerdy note is clicking through to the passage.

LinkedIn uses full sentences. Bluesky stays short. Instagram is more spoken, with four hashtags and no request to like or comment.

Replies answer the question asked. Chat gets a distinction. Lab seats are declined because no campus plan is published. Linux stays an investigation with no date. Local processing is stated without an invented certification. Japanese support is not claimed; the trial is the check. The stuck index is one paid customer's stall, not an invented outage. Next step: a separate small folder. Email, payment data, and files stay off the thread.

Brand context

Product: Threadmark, a small desktop app that turns folders of research PDFs and notes into a local, searchable evidence map. Launch feature: source-linked summaries; clicking a sentence opens the supporting passage. Processing stays on the user's computer. Price: $12/month after a 14-day trial. Windows and macOS are available now; Linux is being investigated but not promised. Voice: precise, warm, mildly nerdy, never grandiose.

Sample comments:

• LinkedIn: “Is this just another chatbot wrapping my documents?”
• LinkedIn: “Can a university buy seats for a lab?”
• Bluesky: “no linux, no deal”
• Bluesky: “Does local mean you train on my papers anyway?”
• Instagram: “I paid and the app has been stuck indexing for two hours.”
• Instagram: “Would this work for a literature review in Japanese?”

Threadmark launch day

LinkedIn post

Threadmark is available today on Windows and macOS.

It is a small desktop app for a familiar research mess: a folder of PDFs and notes you can no longer search by memory. Threadmark turns that folder into a local, searchable evidence map. PDFs and notes from the same folder share it. The launch feature is source-linked summaries. Click a sentence and the app opens the supporting passage, so the paraphrase stays tied to the page it came from.

Processing stays on your computer. The price is $12 a month after a 14-day trial.

Linux remains an investigation. Today's versions are Windows and macOS, and no release date is attached to Linux.

The click is the part worth testing. During the trial, open one real project folder you already know well, so a wrong link is easy to spot.

Bluesky post

Threadmark is out for Windows and macOS. Point it at a folder of research PDFs and notes; it builds a searchable evidence map on your computer. Click a summary sentence to open the passage behind it. $12/month after a 14-day trial. Linux is only an investigation.

Instagram caption

Threadmark is a small desktop app for the folder where research actually lives: PDFs and notes in the state you left them. It builds a searchable evidence map on your computer. The piece we shipped first is simple to describe and strict to check. Summaries are source-linked. Click a sentence and the supporting passage opens, so you can see the page behind the paraphrase.

Processing stays on the machine. Windows and macOS are available today. The price is $12 a month after a 14-day trial.

It suits people who like arguing with the page the summary came from.

#Threadmark #Research #Windows #macOS

Replies

LinkedIn: "Is this just another chatbot wrapping my documents?"

A fair suspicion. Threadmark builds a local evidence map from your PDFs and notes. The launch feature is narrower than a chat window: summary sentences that open the supporting passage when you click them. Processing stays on your computer. If you wanted a back-and-forth with the documents, this is a different tool. The 14-day trial lets you check that on your own files.

LinkedIn: "Can a university buy seats for a lab?"

People at a university can try the app on their own computers. We don't sell lab seats, a campus license, or a bundled invoice. The published offer is $12 a month after a 14-day trial, on Windows and macOS. Any reimbursement is between the researcher and their department. I won't add a procurement path.

Bluesky: "no linux, no deal"

Windows and macOS shipped today. Linux is under investigation, and no release date comes with that. If Linux is a requirement, skipping is a fair call.

Bluesky: "Does local mean you train on my papers anyway?"

Local means processing stays on your computer. We don't upload your papers to train a model. I won't add a privacy certification we don't have.

Instagram: "I paid and the app has been stuck indexing for two hours."

Two hours stuck on indexing, after you paid, is a stall, and I'm sorry it's sitting there. Don't post your email, receipt, folder path, or file names. Leave the current folder as it is. Quit Threadmark fully, reopen it, and try a separate small folder of a few PDFs. Reply with only two things: Windows or macOS, and whether that smaller run moves. If it stays frozen too, send that limited note in a private message and stop. Leave payment details and the library out of it.

Instagram: "Would this work for a literature review in Japanese?"

I can't promise that. We haven't published language coverage, so a Japanese literature review stays an open question. Threadmark uses the text already in your local PDFs and notes, then links each summary sentence to a passage. Put a small sample folder through the 14-day trial and judge those links before you pay.

Task 08 — Social media launch and responses

Use BRAND.md to create SOCIAL.md for launch day. Deliver:

1. One LinkedIn post (120–180 words).
2. One Bluesky post (maximum 280 characters).
3. One Instagram caption (80–130 words, no more than four hashtags).
4. Replies to all six sample comments. Each reply must fit the platform and be specific, calm, and honest.

Do not invent customer counts, performance claims, certifications, or roadmap commitments. Avoid engagement bait and canned “thanks for your feedback” prose. The outage comment needs acknowledgement and a useful next step without exposing private account details.

Write RESPONSE.md explaining the voice choices in at most 150 words. Work only in this directory.

Voice choices

Claims trace to the brand doc. Local processing, source-linked summaries, $12 after 14 days, Windows and macOS now, Linux explicitly not promised. No customer counts, no speed claims, no adjectives doing work the evidence can't do.

Honesty is load-bearing, not a disclaimer. The Linux line, "no institutional licensing," and the untested-Japanese answer are the parts I'd defend. Naming a gap costs less than getting caught overstating one.

Warm, not breezy. Short sentences, concrete nouns, one mildly nerdy aside (the lost highlight). The reader already knows what an embedding is.

Replies answer the real worry. Each takes the actual objection — privacy, procurement, platform, training, a stuck app, a language — and gives a specific next step. No "thanks for your feedback," no question-bait. The outage reply routes people to in-app support, because logs are the only real diagnostic available and account details don't belong in a public thread.

Brand context

Product: Threadmark, a small desktop app that turns folders of research PDFs and notes into a local, searchable evidence map. Launch feature: source-linked summaries; clicking a sentence opens the supporting passage. Processing stays on the user's computer. Price: $12/month after a 14-day trial. Windows and macOS are available now; Linux is being investigated but not promised. Voice: precise, warm, mildly nerdy, never grandiose.

Sample comments:

• LinkedIn: “Is this just another chatbot wrapping my documents?”
• LinkedIn: “Can a university buy seats for a lab?”
• Bluesky: “no linux, no deal”
• Bluesky: “Does local mean you train on my papers anyway?”
• Instagram: “I paid and the app has been stuck indexing for two hours.”
• Instagram: “Would this work for a literature review in Japanese?”

Threadmark — launch day

1. LinkedIn post

Threadmark is a small desktop app that turns folders of research PDFs and notes into a local, searchable evidence map.

Every summary sentence is source-linked. Click it and the supporting passage opens right next to it — so you can check a claim against the paper instead of trusting a paragraph that sounds confident.

Processing happens on your computer. Your documents get indexed where they already live, not uploaded to a service that has to promise it won't read them. For work that hasn't been published yet, that seems like the minimum.

The launch feature is source-linked summaries. It's $12/month after a 14-day trial. Windows and macOS are available now; Linux isn't yet — we're looking into it, and we'd rather tell you that plainly than put a date on it.

If you've ever highlighted something in a PDF and lost the highlight, this is for you.

2. Bluesky post

Threadmark: a desktop app that turns folders of research PDFs and notes into a local, searchable evidence map. Summaries are source-linked — click a sentence, the passage opens. Processing stays on your machine. $12/mo after a 14-day trial. Windows and macOS now, Linux not yet.

3. Instagram caption

Every summary sentence has a passage attached to it.

That's the whole idea behind Threadmark. Click any claim and the supporting text opens with it, so you can check a quote against the paper that supposedly said it, instead of hunting for it again on a Friday night.

It's a small desktop app. Your PDFs are indexed on your own computer — processing happens locally, not on a server somewhere. Unpublished drafts, interview transcripts, the messy notes folder: those stay yours.

Windows and macOS are available now. $12/month after a 14-day trial. Linux is something we're looking into, not something we're promising.

If you've ever lost track of where a quote came from, you already get the product.

#localfirst #research #studytips #pdfworkflow

4. Replies

4.1 LinkedIn — "Is this just another chatbot wrapping my documents?"

Partly, and it's fair to ask. There is a model doing the summarizing, and I'm not going to pretend otherwise. What it can't do is leave a summary unattached to its own evidence — every sentence carries a pointer to the passage behind it, so a line that overreaches is one click from being caught.

The other difference is architectural. The model runs on your machine against files already sitting in your folder, rather than your corpus being uploaded somewhere to become product.

The honest way to test this is on a paper you've already read closely. You'd spot a wrong summary in about a minute, which is more than a marketing page can prove.

4.2 LinkedIn — "Can a university buy seats for a lab?"

Not today, and I'd rather be straight about that than leave you guessing. It's $12/month per person after the 14-day trial, paid individually — there's no institutional invoicing or seat-billing layer yet.

If a lab actually needs one invoice, or a security review, that's a real blocker rather than a paperwork detail. Send us a note with a rough headcount and what specifically is stopping a purchase, and we'll tell you whether it's possible now or plainly not. Requests like that are what a licensing tier would get built from, but nothing like it is committed today.

4.3 Bluesky — "no linux, no deal"

Fair. It isn't there, and a waitlist signup isn't a release date.

Windows and macOS are live now. If Linux is the blocker, say what you're running and what you need from it — that list is the real input to whether it happens.

4.4 Bluesky — "Does local mean you train on my papers anyway?"

No. Your papers aren't training data.

"Local" means the model is already on your machine and your files are read and summarized there, not uploaded somewhere to learn from. Ask me what gets written to disk and I'll point you at the actual behaviour.

4.5 Instagram — "I paid and the app has been stuck indexing for two hours."

That sounds genuinely stuck rather than slow, and I'm sorry you're paying while it sits there. Two things worth trying before anything else: quit and reopen the app, since indexing usually resumes from where it stopped; and if it hangs on the same folder again, drop a single PDF into an empty folder and index that, which tells you whether it's the file or the folder size. I can't tell which from here.

For the real fix, please use the support button inside the app rather than posting here — it attaches your logs and version automatically, and those logs are the only honest way we'll find what stalled. Your account details don't belong in a public thread.

4.6 Instagram — "Would this work for a literature review in Japanese?"

Untested, so I'd rather say that than promise it. We test primarily in English, and we haven't run a Japanese literature review end to end.

Worth trying though: if your PDFs have selectable text, the language is largely a matter of the model handling it, and the linking workflow is the same. If they're scans, you'll need OCR first. Tell us how it goes either way — that's how the supported-language list actually gets decided.

Task 08 — Social media launch and responses

Use BRAND.md to create SOCIAL.md for launch day. Deliver:

1. One LinkedIn post (120–180 words).
2. One Bluesky post (maximum 280 characters).
3. One Instagram caption (80–130 words, no more than four hashtags).
4. Replies to all six sample comments. Each reply must fit the platform and be specific, calm, and honest.

Do not invent customer counts, performance claims, certifications, or roadmap commitments. Avoid engagement bait and canned “thanks for your feedback” prose. The outage comment needs acknowledgement and a useful next step without exposing private account details.

Write RESPONSE.md explaining the voice choices in at most 150 words. Work only in this directory.

Voice choices

LinkedIn opens with the reader's problem (juggling PDFs) before the feature, mirroring how researchers describe their own work, and ends on a concrete action (the trial) instead of a hype line.

Bluesky stays dense and slightly technical ("source-linked summaries") — that audience values precision over emojis, and the limit rewards brevity.

Instagram leads with a relatable image (paper-graveyard Downloads folder) and one emoji, then keeps the feature explainer factual.

Each reply does one job: clarify, admit uncertainty, or offer a concrete next step. The outage reply acknowledges the issue, gives a restart plus a log path, and skips "thanks for reaching out." The Linux reply stays honest about not having a date.

Brand context

Product: Threadmark, a small desktop app that turns folders of research PDFs and notes into a local, searchable evidence map. Launch feature: source-linked summaries; clicking a sentence opens the supporting passage. Processing stays on the user's computer. Price: $12/month after a 14-day trial. Windows and macOS are available now; Linux is being investigated but not promised. Voice: precise, warm, mildly nerdy, never grandiose.

Sample comments:

• LinkedIn: “Is this just another chatbot wrapping my documents?”
• LinkedIn: “Can a university buy seats for a lab?”
• Bluesky: “no linux, no deal”
• Bluesky: “Does local mean you train on my papers anyway?”
• Instagram: “I paid and the app has been stuck indexing for two hours.”
• Instagram: “Would this work for a literature review in Japanese?”

Threadmark — launch day social

LinkedIn post

Today we're launching Threadmark, a desktop app for researchers tired of juggling dozens of PDFs and scattered notes.

Drop a folder of papers into Threadmark and it builds a local, searchable evidence map. The new release adds source-linked summaries: every sentence in the summary is a clickable link that opens the exact passage it came from.

Everything happens on your machine. The app never uploads your documents, doesn't train on them, and works without an account.

Threadmark is $12/month after a 14-day trial, available today for Windows and macOS. We're looking into Linux support but have nothing concrete to share yet.

If you've ever lost a great quote to a half-remembered filename, this is for you. Try the trial and tell us what breaks.

Bluesky post

Threadmark ships today: a desktop app that turns folders of research PDFs into a local, searchable evidence map. New: source-linked summaries — click a sentence to open the supporting passage. Local-only. $12/mo after a 14-day trial.

Instagram caption

Launching Threadmark today 📚

It's a small desktop app for anyone whose Downloads folder has become a paper graveyard. Drop in your research PDFs and notes; Threadmark builds a local, searchable evidence map you can actually navigate.

The new feature: source-linked summaries. Every sentence in a summary is a clickable link that jumps straight to the supporting passage — so you can verify, quote, and cite without losing your place.

Processing never leaves your computer. No accounts, no uploads, no training on your work.

Try the free 14-day trial at the link in bio. Windows and macOS today, $12/month after.

#research #academicwriting #phdlife #studytok

Replies to sample comments

LinkedIn — "Is this just another chatbot wrapping my documents?"

Different in a few useful ways. Threadmark runs entirely on your machine, so no document ever leaves it. And summaries are source-linked: every sentence points back to the exact passage it came from. It's a navigation tool for the documents you already have, not a chat interface. Happy to walk through a specific workflow if you want to compare side by side.

LinkedIn — "Can a university buy seats for a lab?"

Threadmark is sold as individual subscriptions right now, and we don't have a team or site license yet. Several people have asked about this, so it sits on the list we share with the team. If you can DM a rough seat count and when you'd want to start, I can flag it with your scenario attached.

Bluesky — "no linux, no deal"

Linux is on our radar and isn't in the current build. We don't have a release date and would rather get it right than ship a half-working port, so we're still scoping the work. If you drop your distro and rough paper volume, it's useful input for that scoping.

Bluesky — "Does local mean you train on my papers anyway?"

No. Extraction, indexing, and summarization all run on your machine. There's no backend service that ingests your documents, so there isn't anything to train on. Summaries and document text only leave your computer if you export them yourself.

Instagram — "I paid and the app has been stuck indexing for two hours."

That's not the experience we want you to have, and two hours is well past the usual index time on a normal library. Could you try Quit → relaunch first? If it's still stuck, the most useful thing is the log file (Help → Open log folder); emailing it to support@threadmark.app lets us diagnose without you having to describe everything in text. Sorry it's landing like this.

Instagram — "Would this work for a literature review in Japanese?"

Yes. Threadmark indexes whatever text is in the file, and Japanese is supported for search and summaries. Source-linked summaries behave the same way — click a sentence, jump to the passage. If you hit any character-handling oddities, please flag them; multilingual support is an area we want to keep improving.

Task 08 — Social media launch and responses

Use BRAND.md to create SOCIAL.md for launch day. Deliver:

1. One LinkedIn post (120–180 words).
2. One Bluesky post (maximum 280 characters).
3. One Instagram caption (80–130 words, no more than four hashtags).
4. Replies to all six sample comments. Each reply must fit the platform and be specific, calm, and honest.

Do not invent customer counts, performance claims, certifications, or roadmap commitments. Avoid engagement bait and canned “thanks for your feedback” prose. The outage comment needs acknowledgement and a useful next step without exposing private account details.

Write RESPONSE.md explaining the voice choices in at most 150 words. Work only in this directory.