What is the difference between this guide and the Claude Code integration overview?

The integration overview explains what connecting Claude Code to Dailybot does and why it matters. This guide is the hands-on walkthrough with exact commands, configuration files, and troubleshooting steps.

Can I use this in a devcontainer or Docker environment?

Yes. The guide covers container-based setups specifically. Install the CLI in your Dockerfile or devcontainer.json, and ensure authentication persists across container rebuilds.

How do I know the connection is working?

Send a test report using the CLI command shown in Step 5. If it appears in your Dailybot feed within a few seconds, the connection is working correctly.

Connecting Claude Code to Dailybot (step-by-step) - Academy

This is the hands-on guide. If you want to understand why connecting Claude Code to Dailybot matters, read the integration overview. This guide assumes you are ready to set it up and want exact steps.

By the end, Claude Code will automatically send progress reports to your team’s Dailybot feed whenever it completes meaningful work.

What you need before starting

Three things: a Dailybot account (free tier works), a project where Claude Code is active, and terminal access in the environment where Claude Code runs. If you use a devcontainer, SSH remote, or cloud IDE, you will install everything in that environment.

Step 1: Install the Dailybot CLI

Open a terminal in your development environment and install the CLI globally.

npm install -g @dailybot/cli

If you are in a devcontainer, add this to your devcontainer.json features or your Dockerfile so it persists across rebuilds:

{
  "postCreateCommand": "npm install -g @dailybot/cli"
}

Verify the installation:

dailybot --version

Step 2: Authenticate

Run the login command with your email:

dailybot login [email protected]

Dailybot sends a one-time verification code to your email. Enter it when prompted. If your organization has multiple workspaces, you will be asked to select one.

For container environments where interactive login is difficult, you can use an API key instead:

export DAILYBOT_API_KEY="your-api-key-here"

Get your API key from the Dailybot dashboard under Settings > API.

Step 3: Add the reporter script

The reporter script is a lightweight bash wrapper that formats and sends progress reports. Most teams place it in agent_scripts/ at the root of their repository.

The script handles metadata injection automatically — it detects the repository name, current branch, and other context. You provide the message and optionally structured data.

Create the directory and script:

mkdir -p agent_scripts

If you are starting from a template, Dailybot provides a reference script in its documentation. The key function is straightforward: take a message string, attach metadata, and send it to the Dailybot API.

Step 4: Configure Claude Code’s instructions

Claude Code reads instruction files like CLAUDE.md or AGENTS.md to understand how it should behave. Add a section that tells it when and how to report progress.

The critical piece is the reporting threshold. You want Claude Code to report after genuinely significant work, not every minor edit. Define what “significant” means for your team:

Report after: Feature implemented, bug fixed, major refactor completed, plan finished, deployment executed, 3+ related commits.

Do not report after: Typo fix, formatting change, dependency update, uncommitted work, reading or exploring code.

Include the exact command Claude Code should use:

bash agent_scripts/dailybot-report.sh "Description of what was done and why." \
  --metadata '{"model":"claude-code"}'

Step 5: Send a test report

Have Claude Code complete a small task, or send a test report manually:

bash agent_scripts/dailybot-report.sh "Test report — verifying Claude Code connection to Dailybot." \
  --metadata '{"model":"claude-code"}'

Check your Dailybot feed. The report should appear within seconds, showing the message along with metadata (repo name, branch, model).

Step 6: Verify in the dashboard

Open the Dailybot dashboard and navigate to your team’s feed. You should see the test report with a clear label indicating it came from an agent. The metadata shows which model sent it, from which repository and branch.

Troubleshooting

“CLI not authenticated” error: Run dailybot login again, or set DAILYBOT_API_KEY in your environment. In containers, ensure the auth token persists between sessions.

Reports not appearing: Check that your Dailybot workspace has agent reporting enabled. Verify the API key or login session is valid.

Container rebuilds lose auth: Store your DAILYBOT_API_KEY in a .env file that is mounted into the container, or in your docker-compose.yaml environment section. Do not commit API keys to version control.

What happens from here

With the connection active, every time Claude Code completes meaningful work in this project, it sends a standup-style update to your team’s Dailybot feed. Your manager sees agent contributions alongside human check-ins. The work is visible, tracked, and part of your team’s daily rhythm.

Next steps: set up the agent dashboard for monitoring, or connect additional coding agents (Cursor, Codex, Copilot) using similar steps.