Gram

What is Gram?

Gram is a product from Speakeasy for hosting and managing Model Context Protocol servers and the AI agents that use them. You can source tools from existing APIs or write them as TypeScript functions, group them into toolsets and expose them as MCP servers in minutes, with OAuth built in. The same platform now doubles as an AI control plane: it captures sessions from agent tools such as Claude Code, Claude Cowork, Cursor and Codex, stores them, runs risk scanners over the content, and routes MCP traffic through a gateway with role-based access controls, logs, tracing and analytics. A usage explorer breaks consumption down by project, model, department and role. Platform and security teams use it to see what agents are doing and to cap spending. Pay as you go has no platform fee. It charges $0.35 per million tokens for storing sessions, $0.99 per million tokens for risk scanning and $20 per GiB of MCP gateway egress, and Enterprise is priced on tokens under management.

Use cases

• Host MCP servers for your product
• Turn an API spec into agent tools
• Build tools as TypeScript functions
• Scan agent sessions for risk
• Track agent token usage by team
• Add OAuth to an MCP server
• Cap inference spend per key

Who it's for

• Platform engineers
• Security teams
• Developers
• AI platform owners

Why use Gram

• One platform covers building MCP servers and watching how agents use them, so teams avoid stitching separate tools together.
• Billing is metered with no platform fee, which helps small teams start without a contract.
• Usage data can be split by person, role and department, which supports internal chargeback conversations.
• Existing APIs can be turned into tools, so you do not have to hand-build every MCP server.
• Inference caps give a monthly ceiling on platform-run model costs.

Gram features

• Serverless hosting for MCP servers
• Tools sourced from existing APIs or written in TypeScript
• Toolsets that bundle tools into MCP servers
• OAuth, provisioning and role-based access controls
• MCP gateway with egress metering
• Agent session capture from Claude Code, Cursor and Codex
• Risk scanning of agent content with guardrail policies
• Usage explorer by project, model, department and role
• Inference caps set per platform-managed key
• Agents API with chat API keys
• Directory Sync for user and department attributes

Gram integrations

• Claude Code
• Claude Cowork
• Cursor
• Codex
• Stripe

How to get started with Gram

1. Create a project in the Gram dashboard and start the trial.
2. Add tools from an API spec or write them as TypeScript functions with the Gram SDK.
3. Group tools into toolsets and publish them as MCP servers with OAuth.
4. Connect agent platforms such as Claude Code or Cursor so sessions are observed.
5. Use the usage explorer, risk scans and billing page to review activity and set caps.

Gram limitations

• Pay as you go bills three metered products plus pass-through inference at provider cost
• Spend by product estimates are shown only to pay-as-you-go organizations and are not invoices
• SSO, a full audit trail and SLA-backed support require the Enterprise plan
• Cycle totals finalize 72 hours after a billing cycle closes

Gram pricing

Gram has no free plan, but it offers a free trial. Paid plans start at $0.35 usage-based, according to the official pricing page on 24 Oct 2026.

Gram plans

• Pay as you go

  Usage-based

  • $0.35 per million tokens of agent session storage
  • $0.99 per million tokens scanned for risk
  • $20 per GiB of MCP gateway egress
  • No platform fee

• Enterprise

  Custom

  annual contract on tokens under management

  • SSO and full audit trail
  • Early feature access
  • Legal and security reviews
  • Dedicated forward-deployed engineer

Plans as listed on the official pricing page on 24 Oct 2026. Prices can change and may vary by country or exclude tax. Check the official pricing page.

Gram alternatives

The main Gram alternatives listed on Debuly are Composio, Arcade, Speakeasy AI Control Plane, CloudByte AI, Letta and MCP Market. The table compares pricing and platforms; order is A–Z.

1. 

   Arcade

   Actions runtime that lets AI agents call tools in Gmail, Slack, Salesforce and more under each user's own permissions.

   Developer platformsAutomation and agents
   FreemiumVisit Arcade (official site)
2. 

   CloudByte AI

   Observability dashboard for teams using AI coding tools, tracking sessions, token spend, commits, security and DORA metrics.

   Developer platformsCodingProductivity
   FreemiumVisit CloudByte AI (official site)
3. 

   Composio

   Developer platform with SDKs, CLI and MCP that gives AI agents 1000+ pre-authenticated app toolkits and per-user sessions.

   Developer platformsAutomation and agents
   Visit Composio (official site)
4. 

   Letta

   Platform, formerly MemGPT, for building stateful AI agents with persistent memory, from a terminal app, SDK or API.

   Developer platformsAutomation and agentsCoding
   FreemiumVisit Letta (official site)
5. 

   MCP Market

   Directory and hosted hub for MCP servers and agent skills, with toolkits that connect Claude, Cursor and other AI clients to your tools.

   Developer platformsAutomation and agentsCoding
   FreemiumVisit MCP Market (official site)
6. 

   Speakeasy AI Control Plane

   Governance layer between AI agents and company systems that connects MCP servers, filters risky data, scopes access by role and logs every tool call.

   Developer platformsAutomation and agents
   PaidVisit Speakeasy AI Control Plane (official site)

Gram FAQ

Sources & verification

Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.

Image credits

• Logo: from speakeasy.com · fetched 23 Oct 2026 · Vendor trademark, used only to identify the product; removed on request
• Cover: none collected yet, so a generated tile is shown.

Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.