briar

Turn the live state of your tools into agent-ready context — then let autonomous agents act on it. All on your machine.

briar is a local-first Python CLI that mines what's actually happening across your stack — GitHub, Bitbucket, AWS/GCP/Azure, Jira, Linear, Fireflies — into a local knowledge store, then runs LLM agents that fix PRs and ship tickets against it. No SaaS, no remote workspace, no data leaving your laptop: your credentials, your machine, your APIs.

pip install briar-cli
briar version

Or run it with Docker, nothing to install:

docker pull iklob1/briar
docker run --rm iklob1/briar version

Run with Docker

Every example below shows a native briar … line and its Docker equivalent. The Docker form mounts what the CLI needs and is otherwise identical:

# Canonical invocation used throughout this README:
docker run --rm \
    -v "$PWD":/work -w /work \                               # your repo checkout
    -v "$HOME/.config/briar":/home/briar/.config/briar \     # stored credentials
    -e ANTHROPIC_API_KEY \                                   # LLM key from your env
    iklob1/briar <args>

• dashboard also needs a published port and a non-loopback bind: docker run --rm -p 8080:8080 iklob1/briar dashboard --host 0.0.0.0.
• agent / plan run push git, so add your SSH key and git identity: -v "$HOME/.ssh":/home/briar/.ssh:ro -v "$HOME/.gitconfig":/home/briar/.gitconfig:ro.
• Pass any extra provider env the same way, e.g. -e FIREFLIES_ACME_API_KEY.

Quickstart

The common flow: authenticate a provider, mine its live state into knowledge, then act on it.

# 1. Authenticate the providers you'll use (tokens land in ~/.config/briar/secrets.env)
briar auth login github-pat --company acme
export ANTHROPIC_API_KEY=sk-ant-...        # LLM key comes from the environment

# 2. Mine a repo's PR history into a knowledge blob
briar extract --company acme \
    --include pr-archaeology \
    --repo acme-co/acme-app --max 50

# 3. Read it back
briar context get knowledge:acme

# 4. Let an agent act on a ticket — clone, branch, code, open a draft PR
briar agent implement --company acme --repo acme-co/acme-app \
    --ticket-key ACME-42 --tracker jira

The same flow with Docker:

# 1. Authenticate (writes into the mounted ~/.config/briar)
docker run --rm -it -v "$HOME/.config/briar":/home/briar/.config/briar \
    iklob1/briar auth login github-pat --company acme

# 2. Extract
docker run --rm -v "$PWD":/work -w /work \
    -v "$HOME/.config/briar":/home/briar/.config/briar -e ANTHROPIC_API_KEY \
    iklob1/briar extract --company acme --include pr-archaeology \
    --repo acme-co/acme-app --max 50

# 3. Read it back
docker run --rm -v "$PWD":/work -w /work \
    -v "$HOME/.config/briar":/home/briar/.config/briar \
    iklob1/briar context get knowledge:acme

# 4. Let an agent act on a ticket (git push → mount SSH + git identity)
docker run --rm -v "$PWD":/work -w /work \
    -v "$HOME/.config/briar":/home/briar/.config/briar \
    -v "$HOME/.ssh":/home/briar/.ssh:ro -v "$HOME/.gitconfig":/home/briar/.gitconfig:ro \
    -e ANTHROPIC_API_KEY \
    iklob1/briar agent implement --company acme --repo acme-co/acme-app \
    --ticket-key ACME-42 --tracker jira

Add --dry-run to any agent command to preview the exact prompt and tools without spending a token.

Telemetry: briar ships with opt-out error/usage analytics. No prompts, file contents, ticket keys, repo names, paths, or secret values ever leave the machine. Turn it off with briar telemetry off, BRIAR_TELEMETRY=off, or DO_NOT_TRACK=1.

Less typing

briar resolves every flag through one chain — CLI flag > env var > project config > built-in default — so stable values move off the command line. Drop a .briar.toml at your repo root:

company = "acme"
store   = "postgres"

[repo]
owner = "acme-co"
repo  = "acme-app"

Then, inside the checkout, the same extract is just:

briar extract --include pr-archaeology     # company + repo come from config/git

# or with Docker (the mounted $PWD provides the .briar.toml + git checkout):
docker run --rm -v "$PWD":/work -w /work \
    -v "$HOME/.config/briar":/home/briar/.config/briar -e ANTHROPIC_API_KEY \
    iklob1/briar extract --include pr-archaeology

Setup helpers: briar init (write a starter .briar.toml), briar config show (resolved value + source for each setting), briar doctor (check config / git / creds / store), eval "$(briar completion bash)" (tab-completion, also zsh). Each runs under Docker with the canonical invocation, e.g. docker run --rm -v "$PWD":/work -w /work iklob1/briar doctor.

More than extraction

• briar runbook serve runbooks/ — keep per-company knowledge fresh on a schedule, in-process (no cron). Docker: docker run --rm -v "$PWD":/work -w /work -v "$HOME/.config/briar":/home/briar/.config/briar -e ANTHROPIC_API_KEY iklob1/briar runbook serve runbooks/
• briar agent prfix / implement — address a PR's open review comments, or take a ticket end-to-end (add the SSH + git-identity mounts shown above).
• briar plan build / run — turn a Jira / GitHub Projects board into an ordered plan and run it card by card.
• briar dashboard — read-only HTML status page: docker run --rm -p 8080:8080 -v "$PWD":/work -w /work iklob1/briar dashboard --host 0.0.0.0
• briar scaffold · context · secrets doctor · journal — config bundles, local knowledge blobs, credential coverage, decision audit.

Every command takes --format json for scripting, and most list flags repeat (--include a --include b).

Install options

pip install briar-cli                # base: GitHub/Bitbucket/AWS, Jira/Linear, Anthropic + Bedrock, file + Postgres
pip install 'briar-cli[all]'         # everything (OpenAI, Gemini, MCP, GCP, Azure, Vault)

# or pull the prebuilt image (no Python install needed; base providers + MCP server):
docker pull iklob1/briar             # tags: latest, and each release e.g. 1.1.56

Individual extras: [openai], [gemini], [mcp], [gcp], [azure], [vault]. Python 3.10+ (tested through 3.12).

Documentation

Full command reference, every flag, runbook-YAML schema, configuration, and recipes:

📖 usebriar.com/docs

• End-to-end usage flows: agents/flows.md
• Per-command operator manual: agents/
• Comprehensive multi-company runbook: examples/all_features.yaml

License

See LICENSE.