Orchestrate AI agents —
visually or via API
Build teams of specialised AI agents with the drag-and-drop canvas. Or call Claude, GPT-4o, and Gemini through a single OpenAI-compatible gateway. Two interfaces. One platform.
19+
Agent Teams
15+
AI Models
5
Output Types
4
SDKs
20+
Doc Sections
Platform capabilities
One platform for creative users and developers alike.
Visual Agent Builder
Drag-and-drop agent teams with a 3D simulation canvas. Pick roles, connect specialists, and brief your delegation — no code required.
OpenAI-compatible Gateway
One endpoint, every model. Point any OpenAI SDK at our gateway to access Claude, GPT-4o, Gemini, DALL-E, Whisper, and TTS — no SDK swap needed.
Multi-modal Output
Orchestrate text, image, audio, video, and 3D outputs in a single pipeline. Agents hand off between modalities automatically.
Developer CLI & SDKs
The `delegation` CLI covers auth, chat REPL, file ops, usage stats, and project management. Full SDK examples in cURL, Python, TypeScript, and Go.
Google Workspace
Agents connect to Drive, Sheets, Docs, Slides, Forms, and Apps Script via MCP. Read, write, and automate your entire Workspace without leaving the platform.
Human-in-the-Loop
Approve, reject, or redirect at any step. The orchestrator holds before critical actions — you keep the judgement, agents handle the volume.
How it works
Define your team
Pick agent roles from the library or describe what you need. Add a lead and supporting specialists — or start from a pre-built template.
Set your goal
Type a brief. The orchestrator breaks it into tasks, assigns agents, and starts executing with real-time status updates.
Review & ship
Approve steps where you want control. Export text, image, audio, video, or 3D — or integrate outputs directly via the gateway API.
Get an API key
Create a developer account, issue a gateway key from the dashboard, and point your existing OpenAI SDK at our base URL.
Call any model
Switch between Claude, GPT-4o, Gemini, and more with a single model parameter. Streaming, function calling, and embeddings all work out of the box.
Build & monitor
Track usage per key, set rate limits, attach sessions for cost attribution, and pull detailed logs — all from the developer dashboard.
Get started free
One platform.
Two ways to build.
For creators
- Visual drag-and-drop agent builder
- Mix Claude, Gemini, Imagen & more
- Text, image, audio, video & 3D output
- Google Workspace integration via MCP
For developers
- OpenAI-compatible gateway endpoint
- Per-key rate limits & usage tracking
- CLI tool + SDK examples in 4 languages
- Chat session history & audit logs
Ready to delegate?
Start free. No credit card required. Switch to developer mode for direct API access anytime.