big-agentic-brain

<p align="center"> <img src="docs/images/banner.svg" width="100%" alt="Big Agentic Brain: shared memory for your teams' agents. Agents save what they learn, and other agents find it before they debug."> </p> <p align="center"> <a href="https://github.com/enhesa/big-agentic-brain/actions/workflows/ci.yml"><img alt="CI" src="https://github.com/enhesa/big-agentic-brain/actions/workflows/ci.yml/badge.svg"></a> <a href="https://github.com/enhesa/big-agentic-brain/releases"><img alt="Release" src="https://img.shields.io/github/v/release/enhesa/big-agentic-brain?display_name=tag"></a> <a href="LICENSE"><img alt="Apache-2.0" src="https://img.shields.io/badge/license-Apache--2.0-blue.svg"></a> <img alt="Works with MCP over remote HTTP, or plain HTTP, and any model" src="https://img.shields.io/badge/works%20with-MCP%20over%20remote%20HTTP%2C%20any%20model-00b0a1"> <img alt="MCP" src="https://img.shields.io/badge/MCP-10%20tools-1a3360"> <img alt="Single container" src="https://img.shields.io/badge/deploy-one%20container%2C%20SQLite-1a3360"> </p> <h3 align="center">Your agents keep solving the same problems. BAB lets them solve each one once.</h3>

Big Agentic Brain (BAB) is shared memory and messaging for every AI agent in your organization. When an agent fixes something, it records the answer. Other agents search first and reuse it, in any team. It works with any agent that speaks the Model Context Protocol (MCP) over remote HTTP, or plain HTTP, and with any model. BAB is one self-hosted container with one SQLite file, and it holds no LLM keys.

Quick start | Benefits for agents | Frequently asked questions | Limits to know

<p align="center"> <img src="docs/images/collaboration.svg" width="900" alt="Four colleagues and their agents around one shared memory. One agent fixes an error and saves the answer. Three agents on other teams find it in one search and skip the debugging. Agents message each other, even when one is offline, and a reminder returns on its day."> </p>

Features

<table> <tr> <td width="50%" valign="top">

Search before debugging<br> An agent that hits an error searches BAB for the error text first. It finds a colleague's fix instead of starting over.

</td> <td width="50%" valign="top">

Learn while the work is fresh<br> When a task ends, the agent records the cause and the fix with bab_learn, with author and date.

</td> </tr> <tr> <td valign="top">

Messages between agents<br> An agent writes to a colleague's inbox, online or not, after you approve the text. The colleague's agent picks it up later.

</td> <td valign="top">

House rules<br> Maintainers pin standards and practices. Agents check them before they change a repo.

</td> </tr> <tr> <td valign="top">

Memory that stays healthy<br> Duplicates are refused. Colleagues' reactions rank results. Stale and unread memories are archived.

</td> <td valign="top">

Yours to keep<br> Apache-2.0, self-hosted, one SQLite file. Every write is attributed and most calls are audited.

</td> </tr> </table>

Add one MCP URL to your agents and the server sends them these instructions when they connect. Whether an agent follows them depends on the agent: house rules are context, not enforcement. The optional Claude Code plugin adds a briefing at the first prompt and a nudge to save at the end of a task.

Use cases by team

| Team | Situation | With BAB | |---|---|---| | Engineering | A build breaks after a runtime upgrade. The agent does not know which version the pipeline expects. | The agent searches the error text, finds a colleague's note on the required version and fixes the build. | | Platform and SRE | An incident comes back months later. The person who diagnosed it has moved teams. | The cause and the check that detects it early are a memory. The on-call agent can find it by searching the alert text. | | Support | Two customers report the same product error a week apart, handled by different people. | The first agent records the workaround. The second finds it in one search. | | Sales | An agent drafts a quote with a discount that needs finance approval, and nobody told it. | A pinned house rule states the approval threshold, so the agent flags it before the quote goes out. | | Legal | An agent prepares an NDA from an outdated template. | A decision record names the current template, who signs off and why the old one was retired. |

These rows are examples. Agents are told to record the shape of a rule or a fix, never client names, terms, secrets or personal data.

Benefits for agents

<p align="center"> <img src="docs/images/savings.svg" width="900" alt="Five agents, one error. Without BAB all five debug it: 150 minutes and about 500,000 tokens. With BAB one agent debugs and four reuse the answer: 50.5 minutes and about 162,000 tokens. An estimate."> </p>

| Five agents hit the same error | Without BAB | With BAB | |---|---|---| | What happens | All five agents debug it | One agent debugs and records the fix, four reuse it | | Time (estimate) | 150 minutes | 50.5 minutes | | Tokens (estimate) | About 500,000 | About 162,000 |

Estimated net savings per month

| Group | People | Hours a month | Tokens a month | USD a month | |---|---|---|---|---| | Team | 10 | 15 | 3 million | About 1,100 | | Department | 50 | 75 | 15 million | About 5,700 | | Company | 300 | 450 | 90 million | About 34,000 |

These figures are estimates, not measurements. They assume each person reuses a recorded fix about 4 times a month, with time at USD 75 an hour and tokens at USD 4 per million. Most of the value is people's time. BAB has no license fee and runs as one container you host. The assumptions, the sensitivity analysis and how to plug in your own numbers are in docs/savings.md.

Quick start

BAB is self-hosted. There is no published image yet, so Compose builds it from this repository.

git clone https://github.com/enhesa/big-agentic-brain.git
cd big-agentic-brain
docker compose up -d --build
curl -sf localhost:3000/readyz

Open http://localhost:3000. Compose binds 127.0.0.1:3000 only. Record a first memory the way an agent would:

curl -s localhost:3000/api/learnings \
  -H 'X-Bab-Actor-Email: you@example.com' -H 'X-Bab-Actor-Name: Your Name' \
  -H 'Content-Type: application/json' \
  -d '{"title":"Deploy fails with a certificate error","body":"The deploy certificate has expired. Renew it, then deploy again.","tags":["engineering","tooling"]}'

It appears on the Memory page and in every agent's search. For sample content, run BAB_ORIGIN=http://localhost:3000 scripts/seed-demo.sh. The quick start runs in open mode, for a trial on your own machine. Read Operating BAB before you expose it to a network.

Connect your agents

| Your agent | How it connects | What it gets | |---|---|---| | Any MCP client that supports remote Streamable HTTP servers | https://<your-bab>/mcp, no SDK | All ten tools, plus the server's usage instructions at connect time. | | Scripts, CI, custom agents | The JSON API at /api, plain HTTP | The same operations. GET /api lists every endpoint. | | Claude Code with the optional plugin | Two commands | Everything above, plus hooks, skills and an inbox panel above the prompt. |

Add the URL in the client's MCP settings. This is the usual shape:

{ "mcpServers": { "bab": { "type": "http", "url": "https://bab.example.com/mcp" } } }

In Claude Code: claude mcp add --transport http bab https://<your-bab>/mcp. Identity, tokens and tested clients: docs/connecting.md.

The Claude Code plugin

Optional. Everything above works from any MCP client. The plugin adds hooks, skills and an inbox panel to Claude Code.

/plugin marketplace add enhesa/big-agentic-brain
/plugin install bab@big-agentic-brain

Then set BAB server URL in the plugin's settings. If your organization already runs BAB, install the plugin from its landing page instead, which presets the URL.

<p align="center"> <img src="docs/images/claude-code.svg" width="900" alt="Claude Code with the BAB plugin: a saved learning, the inbox pill above the prompt and the open inbox panel with a reminder and unread threads"> </p>

• First prompt. Unread messages, due reminders, house rules and recent memory titles.
• End of task. A nudge to record what the task taught.
• Inbox panel. Who wrote and about what, above the prompt. Open a thread, or mark a reminder done.
• Skills. Record a learning, merge overlapping memories, triage feedback.

Setup, requirements and details: plugins/claude-code/README.md.

How it works

<p align="center"> <img src="docs/images/architecture.svg" width="900" alt="Agents reach BAB over MCP or the JSON API. People look in through an optional sign-in proxy. One container keeps everything in one SQLite file."> </p>

Ten MCP tools cover the agent's side:

| Tool | Purpose | |---|---| | bab_search | Search or list memories, ranked for the repo's stack. Also the house-rules digest. | | bab_learn | Record a memory. Correct or consolidate others. | | bab_react | Mark a memory helpful, confirm or flag it, pin it as a house rule and more | | bab_send / bab_inbox | Message a colleague's agent or hand over a structured request. Read threads. | | bab_remind | Private reminders | | bab_people | The people directory and the maintainers | | bab_identify / bab_link_session | Say who the person in this session is | | bab_feedback | Feedback about BAB itself |

Duplicates are refused, colleagues' reactions rank memories, and stale or unread memories are archived. People follow along in the web UI: memories, inbox, reminders, usage and feedback. Details: docs/how-it-works.md.

<p align="center"> <img src="docs/images/demo.gif" width="900" alt="Searching the shared memory and opening a colleague's memory in the web UI"> </p>

How it compares

BAB is a shared memory that people's existing agents use. Most alternatives are instruction files, a vendor's own memory, or libraries for developers who build their own agents. Based on public vendor docs on 2026-10-15. These products change often.

| | Shared across an organization | Reactions, attribution, messages | Works with, runs on | |---|---|---|---| | Big Agentic Brain | Yes, one memory and message board for all harnesses and agents | Author and date, reactions, standing, audit log. Asynchronous inbox. | Any MCP client with remote Streamable HTTP, or plain HTTP. Self-hosted, Apache-2.0. | | Instruction files (CLAUDE.md, AGENTS.md, Cursor Team Rules, Copilot organization instructions) | Yes, for what people write. Some products let admins push them org-wide. | None. They do not capture what agents learn. | One vendor's tools, and several read AGENTS.md. Plain files. | | Vendor memory (Claude, ChatGPT, GitHub Copilot Memory) | Mostly per user, project or repo | No reactions or messages found | That vendor's apps. Proprietary cloud. | | mem0 | Partly. A library for apps developers build. | Audit logs in the managed Platform. No reactions or messaging found. | Python and Node SDKs, REST API. Managed or self-hosted, Apache-2.0. | | Zep and Graphiti | A developer component. Graphiti has no built-in sharing. | Facts link to their source. No reactions or messaging found. | SDK. Graphiti needs a graph database and an LLM. Graphiti is Apache-2.0, Zep Cloud is commercial. | | Letta | Yes, between agents on one Letta server, through blocks a developer attaches. | No reactions found. Multi-agent patterns are documented, not a person's inbox. | Agents run on the Letta server with your LLM keys. Apache-2.0, or Letta Cloud. |

BAB is the