AI Launch Profile
Model Context Protocol 2026-07-28 specification goes final
The final Model Context Protocol 2026-07-28 specification became the current authoritative protocol version. It replaces connection-level state with self-contained requests, moves capability negotiation to each request and formalizes an extension model that includes Tasks, Skills over MCP and MCP Apps.

At a glance
Launch Snapshot
- Company
- Not publicly confirmed
- Launch date
- July 28, 2026
- Launch type
- Major Update
- Pricing
- MCP is an open protocol specification, not a paid hosted product. Implementations, infrastructure, models, connectors and support can carry separate costs. The official specification and governance pages are the authority for the protocol boundary.
- Free plan
- Yes
- API
- Yes
- Open weights/source
- Yes
Launch Context
Use these links to move from this record into the broader Launch Intelligence database.
Verification & Sources
- Status
- Verified
- Source links
- 4
- Freshness
- Verified July 29, 2026
- Last verified
- July 29, 2026
- Last updated
- July 29, 2026
Key source checks
Suggest a correction
Kingy AI Take
The 2026-07-28 MCP specification makes a consequential architectural trade: stateless, self-contained requests can simplify scaling and recovery, but implementations must now negotiate and validate capabilities on every request. Neutral LF Projects governance improves the standards boundary. Adoption, compatible implementations and security testing—not the document alone—will determine whether the release reduces integration work.
Who it is for
Agent-platform teams, client and server implementers, enterprise architects, coding-tool builders and open-source maintainers that need a dated interoperability target. This record covers the specification release, not the stale existing Kingy tool profile.
What feels promising
Stateless requests, explicit per-request capability headers and a clearer extension model can make mixed client-server deployments easier to scale and reason about.
What feels unproven
Kingy did not run an interoperability suite. Client and server support, migration behavior, authorization, consent and extension compatibility vary by implementation.
Traction notes
A dated final specification gives AI-tool vendors and enterprise platform teams a stable interoperability target. Stateless requests can simplify horizontal scaling and recovery, while explicit per-request capabilities and extensions make optional behavior easier to reason about. The value depends on compatible client and server implementations, not the specification alone.
Source-backed record