Bring your OpenWiki brain. Keep everything.
OpenWiki named the category: wiki memory, a personal brain for agents, built from your connected sources. RepoOps builds the same brain and proves it, with cited freshness, experiential memory, and the security posture it lacks. Adopt your OpenWiki brain in one command; nothing you built is left behind.
Get started
One command adopts your OpenWiki brain
Install RepoOps (free and local), then import the brain you already built. The wiki and connector dumps land as reviewable proposals, so adopting a brain can never write a silently active memory.
# 1. Run RepoOps (free, local, bring your own key) $ npx repoops run # 2. Adopt your existing OpenWiki brain in one command $ npx repoops import openwiki -> wiki pages, code wiki, and connector dumps mapped to proposals # 3. Review and confirm what you want to keep $ npx repoops ask --me "what did we decide about the rollout?" -> a cited answer, sourced from the brain you brought with you
The migration
What carries over, and what you gain
One command, nothing left behind
repoops import openwiki maps an OpenWiki 0.1.0 brain page by page: the personal wiki, the code wiki, and every connector raw dump, each origin-tagged, onto the same import shape every RepoOps connector emits. No copy-paste, no re-fetch, no starting over.
- Outcome: you keep the work you already did instead of rebuilding it.
- How to use: run the import once; the CLI collects your local OpenWiki dir and hands it off.
Imports land as proposals
Every imported page arrives as a reviewable proposal through the same fetch, redact, and write-guard pipeline every connector uses. Nothing external becomes active memory unread, so adopting a brain can never silently write over what you trust.
- Outcome: you decide what becomes active memory, page by page.
- How to use: review the proposals and confirm the sources you rely on.
Ask your brain today
OpenWiki lists full-text search, semantic search, and MCP as planned on its roadmap. RepoOps ships them now: LLM-free BM25, bring-your-own-key vector embeddings with Reciprocal Rank Fusion, and cited answers, plus brain access over MCP so Claude Desktop, Cursor, and any MCP client read the same brain.
- Outcome: search and MCP work on day one, not as a roadmap promise.
- How to use: run repoops ask, or point an MCP client at your brain and query it in place.
A wiki you can prove
OpenWiki re-runs connectors on a schedule, but nothing checks that a page still matches its sources. RepoOps carries per-page verified badges: each page's cited source refs are checked against HEAD, CI checks those citations, and rotten citations lower your Brain Health score.
- Outcome: a page tells you it is fresh as of a specific commit, so you never trust a stale doc.
- How to use: read the verified badge on the Wiki tab before you rely on a page.
Memory that learns and scales
OpenWiki synthesizes what your sources say. RepoOps adds experiential memory, lessons, errors, decisions, and causal links learned from real runs and enforced back into every agent session as pre-merge rules, plus a consent-gated team wiki where members' opted-in slices synthesize into cross-member pages, each attributed to who learned it.
- Outcome: a bug you fixed once is blocked at the merge gate if it tries to ship again.
- How to use: let RepoOps write the lesson; share a slice and the team wiki synthesizes from it.
Your credentials, encrypted
OpenWiki stores connector OAuth tokens in a plaintext ~/.openwiki/.env file. RepoOps keeps connector tokens in an AES-256-GCM encrypted vault, and your brain stays plain files you own, portable over MCP across jobs, machines, and tools. The reverse door does not exist: only RepoOps imports a brain, not the other way around.
- Outcome: your tokens are encrypted at rest, and your memory moves with you.
- How to use: connect a source from My Brain; OAuth sources activate as your operator enables them.
Where to start: download RepoOps (free and local), run repoops import openwiki to adopt the brain you already built, then confirm the proposals you want to keep. Compare the two side by side on the Why RepoOps page, and read the connectors guide to grow your brain from here.