← Back to all sparks
S

Speakeasy

DEVOPS
Velocity10.0

Speakeasy's Gram is shipping daily — multi-MCP chat, Codex hooks, and long-running assistants in one week.

mcp-platformai-assistantsfly-runtimeenterprise-authdeveloper-platform
Current state
Speakeasy's Gram platform is moving at multiple-releases-per-day cadence across two trains. The Platform train has shipped issuer-gated OAuth from the playground, release-stage badges, OpenRouter credit monitoring with auto-reconciliation, a v2 assistant runtime foundation, hook telemetry attribution in Datadog, Codex (OpenAI) hooks support, OTEL forwarding to customer destinations, Slack Block Kit with interactive replies, and a full migration to WorkOS-native auth. The Elements train added multi-MCP server chat configuration with namespaced tool merging, and a resilience fix so a failing MCP server doesn't wipe out tools from healthy ones in the same chat. Long-running assistants gained token-aware context compaction, self-wake triggers, and long-term memory via vector embeddings.
Where it's heading
Gram is being built as an MCP-native assistant platform — every release reads like infrastructure for assistants that compose many MCP servers, run for a long time, recover from failures, and integrate with enterprise auth and telemetry. The architectural choices (multi-MCP merging with namespacing, per-assistant Fly apps, OTEL forwarding, WorkOS) say the target buyer is a platform team building real production agents, not a tinkerer. Self-healing chat history, credit-exhaustion 402 responses, and per-server failure isolation are the kinds of features that only matter at scale — Speakeasy is building for that scale already.
Prediction
Expect Gram to formalize its v2 assistant runtime in the next sprint, add usage-based pricing tied to OpenRouter credits and Fly machine-hours, and ship deeper MCP server lifecycle tooling (version pinning, canary deploys for new tool versions). A managed MCP server catalog is a plausible adjacency given how much of the platform already presumes multi-MCP composition.

Recent moves

  1. 4d ago

    Issuer-gated OAuth from the playground, release-stage badges, and resilient assistant runtimes

    Playground Connect now flows through the issuer-gated OAuth path, Preview and Beta badges show pre-GA features across the dashboard, and assistant runtimes no longer hang after image upgrades. Steady platform hygiene.

  2. 6d ago

    Platform toolset routing and hook telemetry attribution

    Assistant platform toolsets move onto a path-prefixed route in line with MCP servers, and hook telemetry in Datadog becomes filterable by org, project, source, and event. Internal consistency work plus the kind of telemetry attribution that matters when many teams share a Gram tenant.

  3. 6d ago

    OpenRouter credit monitoring, v2 assistant runtime foundation, and MCP server renaming

    OpenRouter credit monitoring and auto-reconciliation land for enterprise orgs, MCP servers can be renamed without breaking URLs, and groundwork for the v2 assistant runtime ships behind the scenes. Cost-visibility plus runtime evolution prep.

  4. 7d ago

    Resilient tool merging across multiple MCP servers

    In multi-MCP chats, a single failing server (e.g. a 401) no longer wipes the merged tool map for healthy servers. A direct follow-on to the multi-MCP support, hardening the failure model.

  5. 7d ago

    Codex hooks, OTEL forwarding, Slack Block Kit, WorkOS-native auth

    ⚡ SPARK

    A major release bundles Codex (OpenAI) hooks support, OTEL forwarding to customer-configured destinations, rich Slack Block Kit interactive replies, and a full migration to WorkOS-native auth. Assistants also gain self-healing chat history and an always-on platform toolset.

  6. 9d ago

    A single chat can now connect to multiple MCP servers

    ⚡ SPARK

    Gram's Elements train adds multi-MCP chat configuration: one chat can connect to several MCP servers, with tools merged and namespaced so identical names don't collide. The architectural primitive Gram needed to be MCP-native at composition time.