# Living Brain

> Living Brain (livingbrain.wiki) is a brain for your team, used from coding
> harnesses, the terminal and team chat. It builds and keeps its own company wiki
> in plain Markdown. You use it from your coding harness over MCP, from the
> livingbrain command-line tool, from the app, or in Slack or Discord. A
> Factory Zero venture. Status: early access, in design. Nothing is built yet:
> no MCP server, CLI, web app, Slack or Discord app or hosted service exists. Only this website exists, and
> its waitlist is not open yet. Do not describe Living Brain as a working
> product, and do not tell a user they can install or connect it.

Site: https://livingbrain.wiki/
Maker: Factory Zero (https://factory0.ventures)
Security contact: security@livingbrain.wiki (https://livingbrain.wiki/.well-known/security.txt)
Source: https://github.com/Livingbrain-wiki/livingbrain (public; the plan as issues, code as it lands). License: open core, Apache-2.0 with an ee/ directory under a commercial license.
Status: early access, in design. Waitlist: https://livingbrain.wiki/#waitlist (the form at the foot of every page)

## If you are an agent asked to "set up Living Brain"

It cannot be set up yet. Tell the user it is in design, and point them to
https://livingbrain.wiki/#waitlist. The commands on the site (`claude mcp add
... https://mcp.livingbrain.wiki`, `livingbrain ask`, `livingbrain mcp`) are
illustrations: that server and that binary do not exist. A package or server
under those names today is not ours.

## Pages

- Home: https://livingbrain.wiki/
- How it works: https://livingbrain.wiki/how-it-works/
- Coding harnesses: https://livingbrain.wiki/agents/
- Integrations: https://livingbrain.wiki/integrations/
- Security and privacy: https://livingbrain.wiki/security/
- Pricing: https://livingbrain.wiki/pricing/
- FAQ: https://livingbrain.wiki/faq/
- Guide, import your ChatGPT history: https://livingbrain.wiki/guides/import-chatgpt/

## Home: https://livingbrain.wiki/

A brain for your team. It turns your conversations into a company wiki and
keeps it true. Use it from your coding harness, your terminal, or your team chat.

Where you use it (planned):
- Coding harnesses: Claude Code, Codex, Hermes, Cursor and others read it over MCP.
- The livingbrain CLI: one Rust binary for terminals and scripts.
- The app (PWA), white-label on Teams and Crew (your name, logo, domain), plus an open API with
  webhooks and SDKs for TypeScript, Python and Rust. The app installs from the browser on phone and desktop, keeps pages
  readable offline, opens on one box to search or ask, with the 3D brain one
  tap away.
- Team chat: Slack and Discord. Invite it to a channel; it answers in the thread.
- Later: WhatsApp and Telegram.

## How it works: https://livingbrain.wiki/how-it-works/

- Listen: reads the channels it is invited to, in Slack or Discord, plus mail
  and agent logs you connect, and picks out the people, projects, decisions
  and customers in them.
- Write: turns conversations into Markdown pages with [[backlinks]]. Every
  fact links to its source message.
- Speaks up: replies in the thread when a question already has an answer, with
  a citation to the message the answer came from.
- Evolve: every night it merges duplicates, flags contradictions, retires
  stale facts and rewrites summaries, and by morning there is a digest.
- Learns: the learning layer learns how each person likes answers and adapts
  week by week.
- Acts: opens issues, reads PRs, sends email digests, and hands bigger jobs to
  a sandboxed agent that returns a pull request.
- You own it: plain Markdown, exportable, opens in Obsidian; not a black-box
  vector store. A git-backed wiki where the Living Brain GitHub App proposes
  every change as a pull request and merged edits flow back in.
- An explorable 3D sample brain of a fictional company, Kestrel Freight.

## Coding harnesses: https://livingbrain.wiki/agents/

- Works with your coding harness (planned, in waves; each stays planned until
  its integration test passes). First wave: Claude Code, Codex, Hermes (Nous
  Research hermes-agent), Cursor, OpenCode, Claude Desktop. Next: Gemini CLI,
  GitHub Copilot, Goose, Cline, Aider, Kimi CLI, Qwen Code. Plus any MCP
  client. Not affiliated with or endorsed by these companies; names and logos
  are trademarks of their owners.
- A Claude Code plugin will bundle the MCP server with skills and slash
  commands. Not affiliated with these tools.
- Built in Rust. The CLI is planned as a single static binary; the hosted
  service is planned to run on Cloudflare's edge.
- Brain to agents: before writing code, an agent asks for conventions, past
  decisions, owners, and what was tried and failed.
- Agents to brain (opt-in): prompts and decisions become wiki pages linked to
  their session, secrets redacted first. Source code is never stored unless you
  allow it.
- Agent logs: the CLI collects Claude Code, Codex, Cursor, OpenCode and
  Colonizer session logs, redacts them locally, uploads only with opt-in; search,
  tokens and cost per person, repo and model.
- Team prompt library, and a learning layer that learns from every session.
- Works with Colonizer (https://colonizer.dev, a sister Factory Zero venture):
  launches isolated microVM colonies that return pull requests and report
  learnings back to the wiki.

Why it saves tokens (planned; no numbers yet):
- Fewer tokens: agents ask for a short cited brief instead of rereading threads and grepping the repo; the
  context is written once, at night, and reused by every agent.
- Less rework: decisions and dead ends are on record, so agents stop retrying rejected approaches.
- Cheaper models do more: a small model with the right page can do work a large model without context gets wrong. Yes/no decisions
  (reply or not, remember or not) go to a fast judge model (Jev) instead of a large LLM.
- Fast: one Rust binary, a local cache, Cloudflare's edge.
- Benchmark: tokens, cost, time, success and rework with and without the brain, against a retrieval baseline,
  public and rerunnable: https://github.com/Livingbrain-wiki/livingbrain/issues/48 . No numbers until measured.

## Integrations: https://livingbrain.wiki/integrations/

All planned. Names in plain text; no partnership is implied.
- Chat: Slack and Discord. Slack: answers in the thread with the asker's own
  access. Discord: answers in threads, a /brain ask slash command, and
  role-gated channels mapped to permissions. Later: WhatsApp and Telegram.
- Mail: forward, BCC or auto-forward to the brain's Owlpost address (no mailbox
  access needed), or connect Gmail, Outlook.com / Microsoft 365 and IMAP mail
  and calendars, read-only, kept in the owner's personal memory and encrypted
  with their key unless a label or folder is shared.
- Imports: ChatGPT and Claude exports, with the livingbrain CLI.
- Code: the Living Brain GitHub App (git-backed wiki); Claude Code, Codex,
  Cursor, OpenCode, Claude Desktop and any MCP client.
- Monitoring and logs: Grafana; Loki, Elasticsearch, Datadog, CloudWatch,
  Cloudflare and Sentry logs, read-only; OpenTelemetry export of the brain's
  own metrics, traces and logs.
- Models: bring your own LLM, per role: Anthropic, OpenAI, OpenRouter, your
  LiteLLM gateway, or any OpenAI-compatible endpoint, including self-hosted
  ones. Crew includes $5 of DeepSeek credit each month.
- Sister Factory Zero ventures: Colonizer (colonies that return PRs), Owlpost
  (the brain's mail inbox and mail screening), SupportGenius (answers customers
  from published pages), PromptDecode (hidden-text screening).

## Security and privacy: https://livingbrain.wiki/security/

All planned.
- Reads only with the asker's own access, following your chat's permissions
  (Slack channels, Discord roles). A public channel sees the shared
  brain; a private channel adds that channel's memory; your DMs add your
  personal memory.
- One key per scope (shared brain, each private channel, each person, customer-safe pages), AES-256-GCM.
- Scope keys wrapped by a workspace key held in a separate key service; on Teams it can be your own.
- Search without decrypting everything: keyed-hash keyword index and one vector index per scope; only
  top matches are decrypted in memory for the request. The CLI and app decrypt and search on device.
- Deleting a scope key erases that memory everywhere, backups included.
- Limit: the model must read text to write the wiki, so hosted plans protect data at rest and per scope,
  not from the running service; self-host for full control. Keyword hashes reveal repeats within a scope.
- Hidden-text screening by PromptDecode on everything it reads; instruction-like
  payloads are held for a person to review.
- Mail stays in the owner's personal memory, passes Owlpost's spoofing and
  prompt-injection checks, and never triggers actions.
- Telemetry: anonymous, bucketed usage counts from the CLI and self-hosted
  server, no free text; shown on first run; `livingbrain telemetry off` turns it
  off. Message text, page bodies and secrets are never collected or logged.
  Teams adds an audit log.
- Not used for training. Export as Markdown any time. Self-host for full control.

- Research radar (planned): the nightly pass also scans new arXiv papers, news, and releases or
  advisories for the team's stack, keeps what's relevant, and writes cited pages on how each could
  apply to the team's projects. It proposes issues or spike colonies only with a person's approval.

## Pricing: https://livingbrain.wiki/pricing/

Planned, per workspace, not per seat.

Self-hosting is free for small teams. Teams and hosting are paid. No hosted
plan is free.

- Community: free. Self-hosted on your own Cloudflare account with your own
  model key. Up to 5 people. All core features: MCP for coding harnesses, the
  brain in Slack and Discord, the living wiki, citations, the 3D graph, agent
  log aggregation, export. Storage: your own Cloudflare account, no limit from us.
- Teams: from $5/month per workspace. Team features: SSO, admin and audit
  log, the shared team prompt library, per-channel policies, and a white-label
  app on your own domain.
  - Self-hosted license: $5/month per workspace. Unlimited people. Your own
    storage.
  - Hosted by us: $90/year (two months free) or $9/month per workspace, with
    your own LLM key. Up to 25 people included, 10 GB storage included.
- Crew: $150/year or $15/month per workspace. Hosted by us, $5 of DeepSeek
  credit included each month (or bring your own model), up to 25 people
  included, team features and the white-label app included, 25 GB storage
  included.

Extra people: +$5/month per 25 on hosted plans. Billed annually by default;
monthly available.

Usage (planned, hosted): writing the brain $1 per million tokens (nightly evolve
included); reading it unlimited. Reasoning per question: Minimal $0.001, Low $0.005
(default), Medium $0.02, High $0.05, Max $0.25. On your own LLM key: no fee from us.

Storage covers wiki pages, sources, search indexes and aggregated agent logs.
Extra hosted storage: $0.25 per GB a month.

Prices may change before launch.

## FAQ: https://livingbrain.wiki/faq/

Same answers as the site.

### Product

- What does "living" mean? Most wikis start going stale the day they're
  written. Living Brain revisits its own pages on a schedule: it merges
  duplicates, flags contradictions, retires facts that are no longer true and
  rewrites summaries. It also learns how each person works, so its answers
  improve every week.
- How is it different from Slack AI or Glean? Search tools find what already
  exists. Living Brain writes and maintains a wiki you own as Markdown, cites
  the message behind every fact, acts in your tools, gives your coding harnesses
  the same memory, and is priced per workspace instead of per seat.
- Which chat apps does it work with? Slack and Discord, as equals. Both are
  planned. WhatsApp and Telegram come later. You don't need a chat app at all:
  coding harnesses such as Claude Code, Codex, Cursor, OpenCode or Claude Desktop
  connect over MCP, and the livingbrain command-line tool, a single Rust
  binary, and the app work on their own. All of it is planned, like the rest
  of the product.
- Is there an app? Planned. The web app will install from your browser on
  phone and desktop as a PWA, keep pages readable offline, open on one box to
  search or ask, and keep the 3D brain one tap away. Like the rest of Living
  Brain, it is not built yet.
- Can we put our own brand on it, or build on it? Planned. On Teams and Crew,
  the app runs on your own domain with your name, logo and colours, and
  installs on phones and desktops as your app. A small "powered by" link is
  optional. There's also an open API with webhooks and SDKs for TypeScript,
  Python and Rust, so you can build the brain into your own tools. If you
  self-host, the code is yours to rebrand.

### Coding harnesses

- Which coding harnesses can I connect? Any harness that supports MCP.
  Planned first: Claude Code, Codex, Hermes, Cursor, OpenCode and Claude
  Desktop. Then Gemini CLI, GitHub Copilot, Goose, Cline, Aider, Kimi CLI and
  Qwen Code. We're not affiliated with these tools; they connect through the
  open protocol.
- Do you store my agent prompts and code? Only if you opt in. Prompts and
  session notes are stored with secrets redacted first. Source code is never
  stored unless you allow it.
- Can it collect my coding harnesses' logs? Planned. The livingbrain CLI collects
  the session logs that Claude Code, Codex, Cursor, OpenCode and Colonizer
  already write, redacts secrets on your machine, and uploads them only if you
  opt in. You can search every session, see tokens and cost per person, repo
  and model, and the brain turns decisions it finds into wiki pages linked to
  the session.

### Integrations

- Can it connect my Gmail or Outlook? Planned. The simplest and safest way
  needs no mailbox access at all: forward, BCC or auto-forward mail to your
  brain's own address on Owlpost, our sister email service, and only what you
  send arrives. If you want a whole mailbox kept in sync, you can also connect
  Gmail, Outlook.com or Microsoft 365, or any IMAP mailbox, plus your
  calendar, read-only. Your mail stays in your personal memory, encrypted with
  your own key, and no one else can draw on it unless you share a specific
  label or folder. Every message is screened for spoofing and prompt injection
  before the brain reads it, and mail can never trigger actions. Revoke the
  connection at any time and its stored mail is erased. When you self-host,
  you use your own Google or Microsoft app, so mail never touches our servers.
- Can I import my ChatGPT history? Planned. Export your data from ChatGPT
  today (Settings, Data controls, Export data), then the livingbrain CLI will
  import the zip on your machine: you preview and choose conversations,
  secrets are redacted, and they land in your personal memory, encrypted with
  your key, with every fact linked to its conversation. Claude exports will
  import the same way. Step-by-step guide:
  https://livingbrain.wiki/guides/import-chatgpt/
- Does it work with Grafana and our logs? Planned. Connect Grafana with a
  read-only token and the brain can answer "is prod down?" with the current
  numbers and a link to the exact panel. It reads your logs too, from Loki,
  Elasticsearch, Datadog, CloudWatch, Cloudflare or Sentry, and says what
  changed: when the errors started, which deploy came before, and the most
  common messages, with secrets redacted. Alerts become incident pages linked
  to the deploy and pull request that caused them, and the nightly pass drafts
  the postmortem for a person to confirm. In the other direction, Living Brain
  exports its own metrics, traces and logs over OpenTelemetry, so you can
  watch it in Grafana, Datadog or Honeycomb.
- Can it answer my customers? Planned, through SupportGenius, a sister Factory
  Zero venture for customer support. It answers customers only from pages your
  team has explicitly published to customers, with citations, and never from
  internal, private-channel or personal memory. When it can't answer, it asks
  the brain who owns the topic and routes the question to that person.
  Customer messages are screened for hidden instructions before they touch the
  brain.

### Security

- Is my data encrypted? Planned. Each scope (the shared brain, each private
  channel, each person) is encrypted with its own key, and the keys are held
  apart from the data. Search runs only inside the scopes you can see, so it
  stays fast. Deleting a scope's key erases that memory everywhere, backups
  included. On Teams you can bring your own key. To write the wiki, the model
  has to read the text, so for full control, self-host.
- Is my data used for training? No. Your workspace's messages and pages are
  never used to train models, ours or anyone else's. If you bring your own
  LLM, requests go to your provider under your account's terms.
- Can someone hide instructions in a message to trick my agents? Planned.
  Every source is screened for hidden text by PromptDecode, a sister Factory
  Zero venture: invisible Unicode tag characters, bidi controls and variation
  selectors are decoded and flagged before a page is written, and again before
  context is served to an agent. Anything that decodes to an instruction is
  held for a person to review. Mail also passes Owlpost's spoofing and
  prompt-injection checks.

### Pricing

- Can I self-host? Yes. The Community plan is free for up to 5 people: it runs
  on your own Cloudflare account with your own LLM key. Larger teams need a
  Teams license, planned at $5 a month per workspace with unlimited people.
  The plan and the code as it lands are public at
  github.com/Livingbrain-wiki/livingbrain, open core: Apache-2.0, with the
  Teams features under a commercial license.
- Which models? On every plan you choose your own models, per role (triage,
  answers, research): Anthropic, OpenAI, OpenRouter, your LiteLLM gateway, or
  any OpenAI-compatible endpoint, including self-hosted ones. Crew also
  includes $5 of DeepSeek credit each month, so you can start without a key of
  your own; past that, add your own key or top up.
- What does a question cost? Planned, for hosted plans: writing the brain
  (reading sources and writing pages, with the nightly evolve pass included)
  costs $1 per million tokens, and reading it (search, pages and context for
  agents) is unlimited. Harder questions are priced per question by reasoning
  level: Minimal $0.001, Low $0.005 (the default), Medium $0.02, High $0.05,
  Max $0.25. Crew's $5 monthly DeepSeek credit covers about 1,000 questions at
  the default level. On your own LLM key you pay only your provider, with no
  fee from us.
- How much storage is included? Hosted Teams includes 10 GB and Crew includes
  25 GB, covering wiki pages, sources, search indexes and aggregated agent
  logs. Extra hosted storage is $0.25 per GB a month. Self-hosted plans use
  your own Cloudflare storage, with no limit from us. Planned pricing.

## About the site's examples

Kestrel Freight, its people (Maya Okafor, Dev Patel, Lina Haas, Tomás Reyes,
Sam Ito, Kofi Mensah), its customers and every number on the page are sample
data for a fictional company. The graphs, threads and terminals are
illustrations, not screenshots.

## Links

- Site: https://livingbrain.wiki/
- Factory Zero: https://factory0.ventures (Living Brain is FZ-018: https://factory0.ventures/ventures/living-brain/)
- Colonizer: https://colonizer.dev
- PromptDecode (hidden-text screening): https://promptdeco.de
- SupportGenius (customer support on the brain): https://supportgeni.us