Deploy Bot

Deploys the bot to Pipecat Cloud using cloud builds.

Parameters

The user specifies the environment as an argument: /deploy-bot dev, /deploy-bot eval, or /deploy-bot prod. If not provided, ask which environment to deploy to.

• dev → config file: deployment/pcc-deploy.dev.toml
• eval → config file: deployment/pcc-deploy.eval.toml
• prod → config file: deployment/pcc-deploy.prod.toml

Steps

1. Deploy to Pipecat Cloud

pipecat cloud deploy --config-file deployment/pcc-deploy.<env>.toml --yes

The --yes flag skips all confirmation prompts. Pipecat Cloud will build the Docker image remotely using the Dockerfile and context specified in the toml file.

2. Verify

Report the result. The deploy command has a 90-second health check window that will often time out with a message like "Deployment did not enter ready state within 90 seconds." This is normal — the deployment was still submitted successfully. Treat this as a successful deploy and report the agent name.

If the build fails, check the build logs using the build ID from the error output:

pipecat cloud build logs <build-id>

Important notes

• Config files are at deployment/pcc-deploy.dev.toml, deployment/pcc-deploy.eval.toml, and deployment/pcc-deploy.prod.toml.
• The .dockerignore at the repo root controls what gets included in the build context.
• Secrets should already be configured on Pipecat Cloud. If the user hasn't set secrets yet, remind them to do so.