Br8n logo

Br8n

br8n captures coding intent, branch state, files, and diffs, then provides resume cards and searchable session history.

Install Br8n

Pick your client, then add it in one click or paste the config.

STDIOPython
~/Library/Application Support/Claude/claude_desktop_config.json
claude_desktop_config.json
{
  "mcpServers": {
    "br8n": {
      "command": "uvx",
      "args": [
        "br8n"
      ],
      "env": {
        "BR8N_BACKEND": "YOUR_VALUE_HERE",
        "AI_GATEWAY_API_KEY": "YOUR_VALUE_HERE",
        "OPENAI_API_KEY": "YOUR_VALUE_HERE",
        "TAVILY_API_KEY": "YOUR_VALUE_HERE"
      }
    }
  }
}

Settings → Developer → Edit Config, merge this into mcpServers, then fully quit and reopen Claude Desktop.

Values are written into the snippet above so you can paste it as is. They never leave your browser.

We couldn’t confirm it starts correctly

We ran the install command but it didn't respond within our test window — this can mean a slow first-time install rather than a real problem.

uvx br8n

Process exited before answering initialize.

This is an experimental automated check and can have false negatives — missing environment variables, a slow cold install, etc. It doesn’t necessarily mean something’s wrong. Last checked 22d ago.

Overview

The br8n MCP server captures coding context at commit boundaries or on demand, including the current branch, open files, diff, cursor position, and a one-line hypothesis. Its Claude Code plugin provides pickup, capture, search, and explore commands backed by a local SQLite journal or an optional Supabase backend. Reach for it when interruptions, meetings, or branch changes make it difficult to recover your working context. Capture and resume work without an account; semantic search and web-based gap filling require embeddings and, for exploration, a Tavily key.

Use cases

  • Capture a debugging hypothesis before joining a meeting
  • Resume work after switching between Git branches
  • Search previous session notes about an implementation decision
  • Explore changed documentation or dependencies after a long absence

Key features

  • Captures branch, files, diff, cursor position, and intent
  • Shows branch-aware resume cards
  • Stores snapshots in local SQLite or cloud Supabase
  • Provides semantic search over captured history
  • Supports optional web-based context exploration
  • Offers local ONNX embeddings without an API key

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked 2d ago

Recent health check succeeded.

13/17 checks healthy over the last 90d

Supply-chain signal

No high-severity advisories surfaced by our automated scan.

Critical 0High 0Medium 0Low 0

Scanned 5d ago via OSV.dev · br8n (PyPI)

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

GitHub stars1Stargazers on the source repository.
Last commit3mo agoMost recent push to the default branch.
Install checkInconclusiveDidn't respond in our test window — often a slow first install.
Directory activity1 viewsConfig copies, upvotes, and views on AllMCPs.

How Br8n works

What br8n MCP server does

The br8n MCP server records the intent behind a development session rather than only preserving repository state. A snapshot can include the active Git branch, open files, cursor position, current diff, and the developer's one-line explanation of what they are trying to solve. That explanation becomes the main heading on a resume card.

The system is designed for interruptions and context switches. After returning to a repository, the pickup command shows the latest hypothesis, recent snapshots, and an indication of how fresh or complete the available context is. This can help reconstruct work after a meeting, a branch change, or a long gap.

How it works

Snapshots are created in two ways: automatically at commit boundaries through an installed post-commit hook, or manually with /br8n:capture. The project does not continuously watch the workspace. Captured records are stored in a searchable journal and associated with the repository and branch context.

The Claude Code plugin exposes these commands:

  • /br8n:pickup displays the current repository and branch resume card.
  • /br8n:capture saves a new snapshot and asks what the developer was working on.
  • /br8n:search <q> queries captured session history.
  • /br8n:explore <topic> starts a gap-filling research flow.
  • /br8n:embeddings reports or changes the active embedding provider.

When coverage is marked as a gap, the explore flow can search the web, retrieve relevant sources, parse them, and add extracted findings to the session knowledge base. Semantic search can use an external embedding key or an on-device model on the local tier.

Setup and configuration

Install the Claude Code plugin with:

text
/plugin marketplace add anthonysuherli/br8n
/plugin install br8n@br8n

After reloading the session, the plugin bootstraps its own environment. The local tier stores data at ~/.br8n/brain.db and serves it through a loopback API at localhost:8002. Set BR8N_BACKEND=local for that mode. The cloud tier uses Supabase and requires BR8N_BACKEND=cloud together with the relevant credentials.

Basic capture and resume require no key. External semantic search can use AI_GATEWAY_API_KEY or OPENAI_API_KEY. The explore pipeline also requires TAVILY_API_KEY. For keyless local semantic search, install the optional br8n[local-embeddings] package; it downloads an approximately 130 MB ONNX model on first use.

Tools and capabilities

The br8n MCP server supports local capture and resume, branch-specific context recovery, searchable snapshots, and optional research for stale context. Its local embedding mode keeps semantic search on the device. A cloud configuration provides authenticated access to the same API for supported companion clients and HTTP-speaking tools.

The resume card reports recent snapshot activity and a coverage level. Search answers are grounded in captured session history rather than general repository history alone. The journal can therefore preserve decisions and hypotheses that are not represented in commits or files.

Limitations and notes

The local tier is intended for one device and does not synchronize data. Cross-machine synchronization, cross-repository search, managed keys, and team sharing are described as future or paid capabilities rather than currently shipped features. The iOS companion consumes the cloud API and uses per-user authentication, so it is not a replacement for the local-only setup.

Only one embedding space is active at a time. Changing providers requires re-embedding because vectors from different models cannot be compared. If br8n detects an accidental provider change caused by a missing API key, it offers the switch instead of applying it silently. Until confirmation, capture and search continue without embeddings and existing data is retained.

The br8n MCP server is most suitable for developers using Claude Code who want lightweight interruption recovery and a journal of their own working hypotheses.

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about Br8n

br8n is an MCP server that connects Claude Code and other HTTP-capable tools to a journal of development context. Its main tools capture snapshots, show branch-aware resume cards, search captured session history, and explore missing context. Data can use local SQLite or an optional cloud backend.

Install it through Claude Code with `/plugin marketplace add anthonysuherli/br8n`, followed by `/plugin install br8n@br8n`. Reload the session afterward; the plugin creates and manages its own environment on first run.

Yes, capture and resume work without an account or key. Local semantic search can also work without an API key by installing `br8n[local-embeddings]`, which uses an on-device ONNX model. External search and web exploration have additional key requirements.

The local backend stores snapshots in `~/.br8n/brain.db` using SQLite and exposes access through localhost. The cloud backend uses hosted Supabase with authenticated access. The local tier is single-device and does not provide synchronization.

Semantic search can use `AI_GATEWAY_API_KEY` or `OPENAI_API_KEY`; local embeddings avoid those keys. The explore or gap-fill pipeline additionally requires `TAVILY_API_KEY` for web search.

Related MCP servers

Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewBr8n AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/br8n?style=directory)](https://allmcps.com/mcp/br8n)
HTML Embed
<a href="https://allmcps.com/mcp/br8n"><img src="https://allmcps.com/api/badge/br8n?style=directory" alt="Br8n on AllMCPs" /></a>