Mastra MCP 2.0 is 2026-07-28 only: which line for Sume?
@mastra/mcp 2.0.0 drops the legacy handshake and speaks MCP 2026-07-28 only. The Sume server defaults to 2025-11-25, so pin the version you test.

@mastra/mcp 2.0.0 is rebuilt on the MCP 2026-07-28 revision and removes the legacy handshake. The Sume MCP server's default protocol version in code is 2025-11-25, so I would not assume the two connect: test 2.0.0 against https://mcp.sume.com/mcp before upgrading, and stay on the release you have already verified until that test passes.
Vendor facts are from the Mastra release notes; Sume facts from the MCP server code and the MCP quickstart, read 2026-10-01.
What does the Mastra 2.0 release say?
The release notes (@mastra/core@1.68.0, September 23) describe @mastra/mcp@2.0.0 as a major rewrite: MCP 2026-07-28 only, legacy handshake removed, suspend/resume replacing the elicitation APIs, and several transport and protocol surfaces removed per the MCP v2 migration guide. They also say there is no initialize handshake or session headers in this revision.
What does the Sume server speak?
The server code sets the default to 2025-11-25 and also accepts 2025-03-26 and 2025-06-18. Its initialize response declares tools.listChanged: false, so it does not push tool-list changes, and names itself sume-remote-mcp. Nothing in the code I read lists 2026-07-28.
| Fact | @mastra/mcp 2.0.0 | Sume server code |
|---|---|---|
| Protocol revision | 2026-07-28 only | 2025-11-25 default; 2025-03-26 and 2025-06-18 also accepted |
| Initialize handshake | Removed | Used, returns serverInfo |
| Tool list change push | Not stated | listChanged: false |
Which @mastra/mcp line should I use?
Keep the line you have already verified against Sume, and run a one-tool check on any upgrade: connect, list tools, call mcp_health. The existing walkthrough, Mastra MCP client with Sume, covers the setup. If a 2.0.0 connection fails, record the error text and the version you pinned, and stay on the working release until the Sume server's supported versions change.
Sources
Related posts
More in Developers
- Nano Banana batch API: Gemini's 24 h batch vs Sume async jobs
Gemini's Batch API trades up to 24 hours of turnaround for higher rate limits. Sume has no batch tier for images: send async or webhook jobs per request.
- OpenAI Agents API durable sessions and Sume job ids
A durable OpenAI Agents API session continues across turns; a Sume render is a separate job. Keep the job id in the session and re-read it, never re-create.
- OpenAI Agents Python MCP error content: reading Sume tool errors
openai-agents-python v0.20.0 keeps MCP error content with structured output. Sume tool failures arrive as isError text with code, message and http_status.
- OpenAI Agents Python MCP non-text content as JSON: Sume results
openai-agents-python v0.20.0 serializes non-text MCP blocks as JSON. Sume tool results are text blocks, with media delivered as URLs inside the JSON.
Written by Sume