Create your Free Voice AI Assistant
Build an agent from a prompt, give it your tools and knowledge, then put it on a phone number, a website, or a bulk campaign.
Documentation · Quickstart · API reference · Community on Discord · Dashboard
npm install @omnidim-ai/sdkimport OmniDimension from "@omnidim-ai/sdk";
const omni = new OmniDimension({ apiKey: process.env.OMNIDIM_API_KEY! });
await omni.calls.dispatch({ agent_id: 123, to_number: "+15551234567" });Same thing in Python:
pip install omnidimensionimport os
from omnidimension import Client
client = Client(api_key=os.environ["OMNIDIM_API_KEY"])
client.call.dispatch_call(agent_id=123, to_number="+15551234567")Get a key at omnidim.io/api-management.
Connect any MCP client to your account and it can create agents, dispatch calls, and read transcripts.
claude mcp add --transport http omnidim https://mcp.omnidim.io/mcp --scope userFor a local process instead, run npx -y @omnidim-ai/mcp-server setup.
Per-client guides.
| Server SDKs | TypeScript · Python |
| Browser | @omnidim-ai/client for in-page voice sessions (guide) |
| Model Context Protocol | Hosted server (OAuth, no install) · Local stdio server |
| Automation | n8n nodes |
| Examples | Omnidim/examples |
| API spec | openapi.yaml |
- Developer documentation and tutorials
- Discord for questions and build help
- Service status
- Bugs in an open repository go to that repository's issues. Anything else: support@omnidim.io
Fixes, documentation improvements, and new examples are welcome. Read the contributing guide in the repository you are changing, and please do not include API keys, call recordings, or customer data in issues or pull requests. Report security concerns privately to support@omnidim.io.
