A personal brain that follows you
Your accounts already hold what you know: email threads, docs, the pages you read, the sessions your agents run. RepoOps turns them into your own cited wiki, steered by your priorities and delivered where you work. Plain files you own, portable across jobs, machines, and tools.
Get started
Three commands to your first cited answer
Install RepoOps (free and local), connect a source to your personal brain, then ask it. Already tried OpenWiki? Adopt that brain in one command.
# 1. Connect a source to your personal brain $ repoops connect --me # 2. Already have a brain elsewhere? Adopt it as proposals $ repoops import openwiki # 3. Ask your brain, cited $ repoops ask --me "what did we decide about the memory rollout?" -> a cited answer, sourced from your own records
What it does
Ingest, synthesize, steer, deliver, and own
Connect your sources
Connectors can feed your personal brain from email, docs, the web, X, Slack, and every agent session you run. Each source lands as reviewable proposals through the same fetch, redact, and write-guard pipeline, so nothing external becomes active memory unread.
- Outcome: context you would have lost to a buried thread is captured for you.
- How to use: connect a source from My Brain; OAuth sources activate as your operator enables them.
Your own cited wiki
RepoOps synthesizes your records into browsable topic pages, people and context, projects, and topics, each citing the records it came from. A page built only from unconfirmed sources wears an unreviewed-sources label until you confirm them.
- Outcome: one place to look instead of ten tabs, with provenance you can trust.
- How to use: review the proposed pages and confirm the sources you rely on.
Steer it with priorities
A priorities document (your version of an instructions file) tells the brain what to track, which topics to follow, and how often. It steers scheduled syncs, web-search queries, and the next synthesis run.
- Outcome: the brain fills with what matters to you, not noise.
- How to use: edit your priorities on the personal-brain page and the next run follows them.
The brain that briefs you
Other brains wait to be read. Yours shows up at session start, ranked by what you have open, alongside the lessons and wiki pages that touch your current files. Proactive delivery, not a dashboard you remember to visit.
- Outcome: the relevant thread or decision arrives before you go looking for it.
- How to use: it rides the existing session briefing; opt out anytime.
Portable over MCP
Your brain is plain files you own, readable from any MCP client. The same memory follows you into Claude Desktop, Cursor, ChatGPT, or whatever tool you code in next, and it moves with you across jobs and machines.
- Outcome: switching tools or laptops does not reset what you know.
- How to use: point an MCP client at your brain and query it in place.
Adopt a brain in one command
Already built a brain in OpenWiki? Bring it in with repoops import openwiki: the wiki and connector dumps map into reviewable proposals, so adopting a brain can never write a silently active memory. The reverse door does not exist.
- Outcome: you keep the work you already did instead of starting over.
- How to use: run the import, then confirm the proposals you want to keep.
Where to start: download RepoOps (free and local), open My Brain to begin, then connect a source and set your priorities. Your first cited answer is three commands away, and everything you add stays plain files you own.