Agents Gateway

A production-ready, MIT-licensed FastAPI service for serving, orchestrating, and managing AI agents — built on Agno. Deploy in one click; scale to your own Kubernetes when you outgrow it.

Quickstart Docs GitHub

What you get

Agent management

Create, configure, version, and serve agents via a REST API. Per-agent prompts, models, toolkits, knowledge.

Teams

Compose agents into teams for coordinated multi-agent workflows. Coordinate or supervisor modes.

Supervisor platform

Job queue, approval flows, containerized workers (Docker & K8s). Production-ready execution.

Toolkits

Pre-built OAuth-managed integrations for Calendar, Email, Contacts, and Drive — Google & Microsoft.

Prompts service

Versioned prompt templates with pluggable storage (PostgreSQL, LangSmith). Backend-aware.

Knowledge base

Per-tenant document indexing on Qdrant. Tag and retrieve at agent or team scope.

Observability

Pluggable OpenTelemetry tracing and logging. Console, OTLP, Sentry, Logtail out of the box.

One-click deploy

Render, Railway, Koyeb buttons. Pre-built image on GHCR (linux/amd64 + arm64). K8s manifests included.

Quickstart

git clone https://github.com/liberzon/agents-gateway
cd agents-gateway

# Start PostgreSQL + Qdrant (seeds demo agents automatically)
docker compose up -d

# Set up Python environment
./scripts/dev_setup.sh && source .venv/bin/activate

# Start the API
./scripts/start_server.sh

Then chat with a demo agent:

export GOOGLE_API_KEY="your-gemini-key"

curl -X POST http://localhost:8000/v2/agents/demo-assistant/chat \
  -H 'Content-Type: application/json' \
  -d '{"message":"What can you help me with?","user_id":"u1","session_id":"s1","timezone":"UTC","locale":"en","stream":false}'

Interactive API docs at http://localhost:8000/docs.

Deploy

Three one-click providers backed by a prebuilt multi-arch image:

Provider	Config
Render	`render.yaml` — image-based
Railway	`railway.toml` — wrapper Dockerfile
Koyeb	`koyeb.yaml` — image-based

Self-hosters can run the Kustomize-based K8s manifests. Cloud-specific manifests for AWS, Azure, and GCP live in agents-gateway/deploy/.

Full per-platform playbook: Deploy guide.

Project links

Source code — github.com/liberzon/agents-gateway
Deploy artifacts — github.com/liberzon/agents-gateway-deploy
Container image — ghcr.io/liberzon/agents-gateway:latest
License — MIT
Issues — Report a bug

FAQ

Is it production-ready?

Yes, with caveats. The gateway, supervisor, queue, and toolkits are covered by an integration test suite that runs against real Postgres/Qdrant on every push. Pick your own observability and harden the auth before shipping to users.

How does it relate to Agno?

Agents Gateway is a service layer on top of Agno. Agno provides the agent primitives (LLM calls, tools, sessions); the gateway provides the HTTP API, multi-tenancy, persistence, supervisor orchestration, and operational surface.

What models does it support?

Anything Agno supports — OpenAI, Anthropic, Gemini, plus other providers. Configure per agent via the model field.

Does it need a vector DB?

Qdrant is bundled via docker-compose for the knowledge base. If you don’t use knowledge entries, you can run without it. The chat path doesn’t depend on Qdrant.

Can I run it without docker?

Yes — you need Postgres 14+ and (optionally) Qdrant reachable over the network. See Deploy.