Every agent, one gateway
Stop pasting the same server config and API key into Claude, then Cursor, then Codex. Set up and authenticate a server once in Toolport and it shows up in every agent, with hot toggles and no restarts.
All your MCP servers. One port.
Set up an MCP server once (the connectors that give your AI real tools, like GitHub, Slack, or a database) and every agent you use shares it, whether that's Claude, Cursor, or Codex. Tools load on demand instead of dumping hundreds into context, so your agent stays fast and cheap. Every call is checked for tampering, and your keys stay on your machine. New: write your agent rules once and Toolport applies them to every client, instead of hand-syncing CLAUDE.md, AGENTS.md and GEMINI.md.
Windows, macOS & Linux · no account, no cloud
Setting this up for a team? Toolport for Teams →
Works with Claude · Cursor · VS Code · Windsurf · Codex · Antigravity · See all supported agents
how it works
Hundreds of tools across your servers. A handful in your agent's context.
- 1Install Toolport. No account, no cloud.
- 2Add each MCP server once, or import the ones your agents already have.
- 3Toggle it on in every agent. No restarts.
Stop paying the tool tax
Tool definitions, counted in tokens, load on every request before you've typed a word. Fewer tokens mean sharper answers, more headroom under your plan's limits, and lower bills.
Without
Toolport 23,698tokens
Toolport 23,698tokens
With
Toolport 886tokens
Toolport 886tokens
Those are tool-definition tokens on one request. Graded at the same task success, Toolport uses 74 to 91% fewer across servers · reproduce the benchmark
Tool-integrity checks
Toolport fingerprints every tool and flags rug pulls (a definition changing after you approved it) and tool poisoning (hidden instructions in a description). On by default, entirely local.
Secrets in the keychain
API keys live in your OS keychain, injected at runtime. Never in a config file, never in the cloud.
Per-tool governance
Toggle any tool on or off. One switch hides every destructive tool from every agent, fleet-wide.
Context stays flat as you scale
Instead of every server dumping its full tool list into context, Toolport exposes a handful of meta-tools the agent searches on demand. Measured on a frontier model: up to 91% fewer tokens at the same task success, graded for correct answers.
Live observability
Per-server latency, error rates and a full audit trail of every tool call, built into the app.
beyond servers
Your rules and guardrails, on every agent too.
Toolport manages more than MCP servers. The same app keeps your agent instructions and your permission policy consistent across every client on your machine.
Agent rules: write them once
Import your existing CLAUDE.md or AGENTS.md, or write a ruleset in Toolport, and apply it to every client you switch on: Claude Code, Codex, Gemini CLI, Zed and ten more. Works globally and per project, your own text is never touched, and a block you edit by hand is never silently overwritten.
Agent permissions: rules with teeth
One policy in Claude Code's own rule syntax (never delete recursively, ask before any push, never read .env) written into Claude Code itself and enforced in Cursor through a guard hook. Covers shell commands, file reads and MCP calls, with one-click presets.
60 seconds, real app: add a server · every agent shares it · lazy discovery · deny actually blocks
Toolport for Teams
Running agents across a team?
One shared, governed set of MCP servers for your whole team, with each person's keys staying on their own machine, free for up to 5 people.
Get Toolport for TeamsEvery MCP server in one place, without the token tax.
Set them up once and every AI agent you use shares them, loading a handful of Toolport meta-tools instead of hundreds.
Download Toolport