ToolsEnabledSERVICES

ToolsEnabled Services

Make your product
usable by AI agents.

ChatGPT and Dots, Claude, Codex and Cursor: built, security reviewed and maintained. One MCP server, connected to your own user accounts.

Book a 30-minute callSee what you get

Fixed-scope proposal/Independent security review/You publish under your name

01Why now

Agents use
clean, safe tools.

  • DOTS

    OpenAI’s DevDay (Sept 29, 2026) launched Dots: always-on ChatGPT agents that act for a user in the background. It also introduced MCP Events (draft specification), which let a user tell an agent “watch for X in your product and do Y when it happens.”

  • TOOLS

    Agents can only use products that expose clean, safe tools. Products without them get skipped, or get a fragile community hack instead.

02What you get

Pilot or Full.
Fixed price.

One standard MCP server for your product works across ChatGPT, Claude, Codex and Cursor.

Pilot · read-only

$5,000 fixed

about 2–3 weeks

  • MCP server for your product (Streamable HTTP), up to ~5 read tools
  • Sign-in with your user accounts (OAuth 2.1 + PKCE), no shared keys
  • Independent security review before launch
  • OpenAI submission package (you publish under your verified org): test cases, reviewer account plan, privacy text, demo script
  • Works with Claude, Codex and Cursor

MCP Events are part of Full.

Full · read + act

$12,000 fixed

about 4–5 weeks

  • MCP server for your product (Streamable HTTP): read and write tools with clear approvals
  • Sign-in with your user accounts (OAuth 2.1 + PKCE), no shared keys
  • MCP Events (“watch and act”: alerts, new items, status changes)
  • Independent security review before launch
  • OpenAI submission package (you publish under your verified org): test cases, reviewer account plan, privacy text, demo script
  • Works with Claude, Codex and Cursor

Founding customers: the first 3 Pilot customers pay $3,000 instead of $5,000, in exchange for a public case study.

Maintenance: $750/month. It covers hosting (or deployment to your infrastructure), monitoring, updates when MCP or OpenAI policies change, re-submissions, and security patches.

Prices are estimates for typical APIs; the final scope is fixed in the proposal.

03How we work

Five steps.
No surprises.

From a first call to a published integration, with the scope fixed before the build starts.

Book a 30-minute call
  1. A 30-minute call

    Your API, your users, and the 3 things they’d most want an agent to do.

  2. A fixed-scope proposal

    The tool list with read and write labels, the event list, and a data inventory.

  3. Build, with tests written first
  4. An independent security review

    Covering authorization, cross-tenant isolation, token handling and injection.

  5. Submission

    You publish under your name, and we handle reviewer follow-ups.

04Who we are

We build
for agents ourselves.

ToolsEnabled, Inc. builds, security-reviews and maintains the connection between your product and AI agents.

Fleet

ToolsEnabled Fleet

An open-source MCP server that runs Codex and Claude Code agents as one team in NVIDIA OpenShell sandboxes. It’s in public beta.

Bench

ToolsEnabled Bench

An open-source benchmark builder.

Connector

Our own connector

We built and security-reviewed a ChatGPT/Dots connector for our own product on our existing account system, using OAuth 2.1 and read-only MCP tools; it launches soon.

Our open-source products are MIT licensed: github.com/ToolsEnabled.

Make your product
agent-ready.

Book a 30-minute call

Joshua Pinckard, founder · josh@toolsenabled.ai
ToolsEnabled is not affiliated with or endorsed by OpenAI, Anthropic, Cursor or NVIDIA. Product names are trademarks of their owners.