Your First 10 Minutes
This page is a condensed, task-oriented path. It covers four things: install, first session, first edit, and enough of the mental model to keep going. Each step links to the full reference page if you need more detail.
1. Install
Section titled “1. Install”go install github.com/instagrim-dev/bmo@<release-tag>Replace <release-tag> with the reviewed version from the GitHub release page
(for example vX.Y.Z).
You need Go 1.26 or later. The binary goes into $GOPATH/bin (usually
~/go/bin). Make sure that directory is on your $PATH.
Full details: Installation
2. Set an API key
Section titled “2. Set an API key”Export a provider key before starting BMO:
export ANTHROPIC_API_KEY=sk-ant-...BMO supports multiple providers (OpenAI, Gemini, Bedrock, local backends, and others). See Authentication for the full list.
If this fails: confirm the environment variable is set in the same terminal
where you run BMO (test -n "$ANTHROPIC_API_KEY" && echo configured) and
switch to a model that matches the provider you configured.
3. Start a session
Section titled “3. Start a session”cd your-projectbmoBMO opens its interactive TUI. On first run, type /init to have BMO analyze
your codebase and write a project context file (AGENTS.md). This improves
response quality in future sessions.
What you should see: the TUI status line should show an active session, the selected model/provider, and a prompt box ready for input. If provider authentication fails, the run result should show the provider error and the sidebar should surface the affected provider.
4. Ask a question
Section titled “4. Ask a question”Try something concrete:
What does this project do?Where is the main entry point?BMO reads your files, assembles context, and answers based on the actual code.
5. Make a safe first edit
Section titled “5. Make a safe first edit”Start with a small reversible workspace change before asking BMO to touch production code:
Create scratch/bmo-first-edit.md with a one-paragraph summary of what this project doesBMO shows the proposed change and asks for confirmation before writing to disk. You and the agent share the same workspace — there is no separate staging area or hidden branch.
6. Useful commands to know
Section titled “6. Useful commands to know”| Command | What it does |
|---|---|
/new | Start a fresh session |
/model | Switch AI model |
/mode | Change session mode (code, chat, review, debug, plan, sre) |
/help | Open the command palette |
/quit | Exit BMO |
Full list: Quickstart — Essential slash commands
7. Understand the model
Section titled “7. Understand the model”BMO is not a chat wrapper. A few things that matter early:
- Shared workspace. You and the agent operate against the same files, session state, and data directory. Edits are real edits.
- Tools are explicit. File reads, shell commands, LSP actions, and everything else happen through named capabilities that you can inspect.
- Modes shape behavior. Session modes (code, review, plan, debug, sre) change how BMO prioritizes tools and structures responses.
For the full mental model, read How BMO Thinks and Acts.
What to do next
Section titled “What to do next”You now have a working setup and a basic understanding of the system. From here, explore the reference sections below based on what you need:
- Common Workflows — practical patterns for daily coding tasks
- Workflow Map — visual overview of how requests turn into tool use and automation
- Configuration — customize BMO for your project
- Tools Reference — the full list of agent capabilities
Contributing to BMO itself? Start with
CONTRIBUTING.md
for the first local build, starter lanes, and validation commands.