Before you start
ABOUT-ME turns a short questionnaire into a context file that AI tools can read at the start of a conversation. You answer the questions once. The tool writes a full Markdown file, a JSON file that holds your answers and both outputs, and a short block you can paste into any chat.
You need one thing: an Anthropic API key. Create one at console.anthropic.com. The key is your own and is billed to your own account.
Build your file
Add your API key
Paste it into Your Anthropic API key at the top of the form. Click Show to check it. The model list loads on its own once the key is in.
Pick a model
The dropdown lists every model your key can use, pulled live from Anthropic, newest first. The tool starts on the newest Sonnet. Use Refresh to reload the list. If the live list can't load, a built-in list appears instead, with a note saying so.
Answer the questions
Start with who you are, then choose your creative field. Pick the one that fits closest. A set of field-specific questions appears under it. The last section covers your clients and current work. Everything autosaves as you type.
Check the input estimate
Above the button, Input ≈ … tokens shows roughly how large your request is. Each long answer is capped at 600 characters and each short one at 200, so the request stays small.
Build
Click Build My ABOUT-ME Files. Text appears in the Markdown box as it is written. A build takes about a minute, depending on the model. When it finishes, all three outputs appear and the file joins your recent list.
The three outputs
| Output | What it is | Use it for |
|---|---|---|
| ABOUT-ME .md | The full context file in Markdown: seven sections with ## headings, about 400–600 words. | Long projects, custom instructions, or a project file where you want the AI to know you well. |
| ABOUT-ME .json | Your answers, the model used, and both outputs in one structured file. | Backing up your work, reloading it into the tool later, or feeding it to another tool. |
| Short block | 150–200 words of plain text with no formatting. | Pasting at the start of a quick conversation. |
Each output has its own Copy button. The .md and .json have download buttons. The .json file is named with your name and the date, for example ABOUT-ME-jordan-wren-2026-10-17.json.
Edit and save
The Markdown and short block are editable. Type in either box and a Save edits button appears. Saving updates the file in your recent list, and the JSON updates with it.
Word counts sit under each box. The Markdown targets 400–700 words and the short block 150–200. A count turns amber when it falls outside its range. That is a nudge, not an error.
Regenerate the short block
If the short block misses the mark but the full file is right, click Regenerate short. The tool rewrites only the short block, using your full file as the source.
Recent files
The last five builds are kept on this computer. Each one lists the name, field, date, and model.
- Load restores the form answers and all outputs. Tweak anything, then rebuild to save a new version.
- .json and .md download that file directly.
- Delete removes it from this computer. It cannot be undone.
Only the newest five are kept. The sixth build pushes out the oldest one, so download anything you want to keep long term.
Profiles and import
Your answers autosave to this browser, so they are still there when you return. Three buttons under the form manage them:
- Export profile saves your answers as a JSON file. Keep it as a backup, or move it to another computer.
- Import profile loads a JSON file made by this tool. If the file includes outputs, they appear too. Files from other tools or other versions are refused with a reason.
- Clear form empties every answer. Your recent files are not affected.
Keys and privacy
| Setting | Where the key is kept | How long |
|---|---|---|
| Remember my key (on) | This browser's local storage | Until you click Forget key |
| Remember my key (off) | This browser tab only | Until you close the tab |
The key is sent only to api.anthropic.com. Your answers are sent only to that same address, as part of a build request. The tool has no server of its own.
Troubleshooting
The message says the key was rejected (401).
Check the key for missing characters or spaces. Keys start with sk-ant-. Make sure it is active in your Anthropic console.
The model list shows a warning about a built-in list.
The live list could not load, usually because there is no key yet, the connection dropped, or the key was rejected. The built-in list still works for building. Click Refresh once the problem is fixed.
A build says the output hit the token limit.
The file is longer than the tool allows in one go. Shorten two or three of your longest answers, then rebuild. The end of the file is the part that gets cut off.
A build says the response was incomplete.
The connection likely dropped mid-build. Click Build again. Nothing is saved from a partial build.
The rate limit message (429) or overloaded message (529) appears.
Wait a minute and try again. These come from Anthropic, not from the tool. Trying a different model can also help.
I can't save my profile.
The browser is blocking local storage, often in a private window. The status line under the form says so. Use Export profile to keep a copy.
Import says the file is not an ABOUT-ME profile.
Only files made by this tool are accepted. Use the file from Export profile, or the .json from a build.
FAQ
Which model should I pick?
The newest Sonnet is the default and a good balance of quality and cost. Opus writes with more depth at a higher price. Haiku is faster and cheaper, and works well for a short block.
Who pays for a build?
Your Anthropic account, through the key you added. The tool does not charge anything.
Can I use the files with other AI tools?
Yes. The Markdown works as a project or custom instruction file. The short block works in any chat. The JSON is plain structured data.
Why are my answers capped in length?
Long answers make the request slower and more expensive, and rarely improve the result. 600 characters is about three sentences.
Does the tool invent details about me?
The instructions tell the model not to. Read the output before you use it, and edit anything that doesn't sound right.
Where is my data stored?
In this browser: your profile, the recent files, and the model choice. Nothing is stored on a server.