​
Prerequisites

  • Node.js 20, 22, 24, or 25 (enforced by package.json engines)
  • Yarn — the repo uses Yarn Berry with a committed .yarnrc.yml
  • MongoDB — connection string for MONGO_URI
  • Python 3 with uv — only if you are running the voice agent worker

​
Install and run

1

Install dependencies

2

Create your environment file

At minimum set MONGO_URI, APP_URL, and one LLM provider key (ANTHROPIC_API_KEY or OPENAI_API_KEY) with the matching LLM_PROVIDER and LLM_MODEL. Integration credentials can wait until you connect a service.
3

Start the server

The API comes up on http://127.0.0.1:3000. An interactive API explorer is served at /explorer.

​
Authenticate

GoConva accepts two credential types.
Log in to exchange email and password for a token:
Send it on subsequent requests:
Use this for anything acting on behalf of a signed-in person.
API keys grant workspace-wide access. Store them in a secret manager, never in client-side code or a committed .env.

​
Verify it works

A JSON response means the server is up. From here, connect an integration or read the core concepts.

​
Running the docs site

The documentation you are reading lives in src/docs and is built with Mintlify:
It serves on http://127.0.0.1:3001. If the dev server behaves oddly after a large edit, clear its cache with ./.mintlify-clear.sh.