Setup

This page gets Coda running locally and connected to a GitHub repo.

Prerequisites

Docker Desktop (install)
OpenAI API key (get one)
GitHub account (for the Personal Access Token in Step 2)

Manual
Agentic

Clone Coda

Click Use this template on the Coda repo to create your own, then clone it. Or clone directly:

git clone https://github.com/agno-agi/coda && cd coda

Configure API keys

cp example.env .env

Set OPENAI_API_KEY and GITHUB_ACCESS_TOKEN in .env.The GitHub token must be a Fine-grained Personal Access Token with these permissions on the repos you want Coda to work with:

Permission	Access
Contents	Read and write
Pull requests	Read and write
Issues	Read and write
Metadata	Read-only

Full setup: GITHUB_ACCESS.md. Use a Fine-grained token. Classic PATs lack the scoping Coda relies on.

Tell Coda which repos to learn

Edit repos.yaml:

repos:
  - url: https://github.com/your-org/your-repo
    branch: main

Repos in repos.yaml must also be granted in your GitHub token’s repository selection. If you list a repo Coda can’t access, the clone fails silently.

For your first run, point Coda at the agno repo. It gives you something concrete to ask questions about while you get set up.

Start Coda

docker compose up -d --build

First run pulls the base image, builds Coda, and clones the repos in repos.yaml into /repos/. Takes a minute or two.

Verify it's running

curl http://localhost:8000/health

Expect {"status":"ok"}. Connection refused means the container is still starting. Check docker compose logs -f coda-api and wait for the Agno banner.

If you have a coding agent like Claude Code or Codex, paste this prompt into a fresh session in an empty directory. The agent clones the repo, configures it, starts Coda, and verifies it’s running.The prompt below never asks the agent to read or transmit your API keys. The agent opens .env and repos.yaml for you to edit, then waits for you to confirm before continuing.

Set up Coda locally for me.

Steps:
Clone https://github.com/agno-agi/coda into the current directory and cd into it.
Copy example.env to .env. Open .env in the user's editor (or print the path and tell them to open it). Tell the user: "Set OPENAI_API_KEY and GITHUB_ACCESS_TOKEN in .env, save the file, then tell me to continue. The GitHub token must be a Fine-grained PAT with Contents R/W, Pull requests R/W, Issues R/W, Metadata R on the repos you want Coda to know about. See docs/GITHUB_ACCESS.md for the walkthrough." Wait for confirmation before proceeding. Do not ask the user to paste either key into this chat. Do not read the values from .env or print them at any point.
Open repos.yaml in the user's editor. Tell the user: "Add the GitHub repos you want Coda to learn (URL and branch). For your first run, you can leave the default agno repo. Save the file, then tell me to continue." Wait for confirmation.
Check that Docker Desktop is running (`docker info`). If not, tell me to start it and stop.
Run `docker compose up -d --build`. Stream the output so I can see progress.
Poll `curl -s http://localhost:8000/health` every 3 seconds for up to 90 seconds. Proceed once it returns `{"status":"ok"}`.
Print next steps: open os.agno.com, Add OS → Local → http://localhost:8000.

If any step fails, stop and show me the error. Don't try to fix it on your own.

Connect to AgentOS UI

Open os.agno.com and log in. Click Add OS, choose Local, enter http://localhost:8000, click Connect. Try a question:

“What repos are available?”

“Walk me through the auth flow in agno.”

“Where are MCP tools registered?”

“What changed in the last 5 commits on agno?”

Coda routes through its team. The Explorer reads code and traces call chains. The Coder writes in isolated worktrees. The Planner breaks features into issues. The Researcher searches the web. The Triager labels and closes issues.

What you have now

Component	Detail
Coda team	Explorer, Coder, Planner, Researcher, Triager. The Leader coordinates.
Repos	Cloned to `/repos/<name>/` inside the container. Synced every 5 minutes.
Knowledge	Conventions and patterns Coda discovers, stored via Agno’s Learning Machine.
Memory	User context held by the Leader (who you are, what you’re working on).
Scheduled tasks	Repo sync runs every 5 minutes. Daily digest and issue triage activate when their channels are configured.

Connect to Slack →

Welcome

Tutorials

Runtime

Demo OS

Prerequisites

Connect to AgentOS UI

What you have now

Next

​Prerequisites

​Connect to AgentOS UI

​What you have now

​Next

Prerequisites

Connect to AgentOS UI

What you have now

Next