A2A agent card: what Sume ships instead (MCP, Agent Completions)
Sume does not publish an A2A agent card. It has a hosted MCP endpoint and Agent Completions with one model, sume-agent. Which door fits, and when.
An A2A agent card is a structured description an agent publishes so other agents can read its capabilities and contact methods. Sume's docs do not describe one. What Sume documents are two other doors: a hosted MCP endpoint at https://mcp.sume.com/mcp, and Agent Completions, which lets you call the Sume Studio Agent like a model.
A2A facts are from its 2026-08-27 blog post; Sume facts from MCP and Agent Completions, read 2026-10-01.
What is an A2A agent card?
The A2A post says A2A was accepted as a Growth Stage project at the Agentic AI Foundation. It describes MCP as the vertical layer connecting agents to tools and databases, and A2A as the horizontal protocol for peer-to-peer collaboration, where an agent can publish an "agent card" detailing its capabilities and contact methods.
What does Sume offer instead?
| Door | Shape | Use it when |
|---|---|---|
| Hosted MCP | Tools at https://mcp.sume.com/mcp for remote MCP clients | Your agent should call Sume tools itself |
| Agent Completions | POST an instruction or messages, get an async receipt with status_url and cancel_url | You want the Studio Agent to run the whole task |
How does Agent Completions behave from another agent?
The model field accepts only sume-agent, and omitting it gives the same agent. generation_spend_cap_usd is required; omit it and the request fails with 400 invalid_request. The reply is a receipt with status queued, then you poll status_url or cancel through cancel_url. Statuses are queued, processing, completed, failed, canceled. See Agent Completions as a video agent API.
Should I wait for A2A support?
The docs promise nothing about A2A, so do not plan around it. A caller that speaks MCP can use the hosted endpoint today; one that does not can use the HTTP receipt flow. For a wider comparison, read MCP vs CLI vs API for AI agents.
Sources
Related posts
More in Developers
- Captions misspell your brand name: fix it with authored cues
Sume video captions have no custom dictionary field. Pass cues with your own spelling and speech-to-text is skipped, so the brand name prints as written.
- AI Act Article 50(3): what Sume video inspect returns
Video inspect returns probe facts, stills and optional STT. The FAQ ties Article 50(3) to emotion recognition; the inspect docs list no such output.
- EU AI Content Code: Section 1 providers or Section 2 deployers?
Section 1 of the EU Code is for providers marking AI output; Section 2 is for deployers labelling deepfakes. A team shipping API video starts with its own role.
- AI music exactly 30 seconds: no duration field, use Timeline
Music Router rejects duration and duration_seconds. Ask for the length in the prompt, then fix the exact output length with Timeline 1.0 audio.duration_seconds.
Written by Sume