<MCP.Playground/>
online
MCP Agent Studio

Connect Any MCP Server.

MCP Agent Studio is a browser-based MCP client that runs any hosted or remote MCP server with Claude, GPT-5, Gemini, DeepSeek, Grok and 40+ other AI models — side by side, with every tool call traced live. No install, no SDK, no API key.

Sign in with email

Free credits on sign-up · No credit card · Terms

Or browse agent templates · hosted MCPs — no sign-in needed.

Discover the best, most cost-effective model for your AI agent use case

ClaudeClaude
OpenAIOpenAI
GeminiGemini
DeepSeekDeepSeek
MistralMistral
GrokGrok
QwenQwen
NVIDIANVIDIA
MiniMaxMiniMax
KimiKimi
+5
more

What you can do

One workspace, four ways to ship with MCP

Chat with your MCP

Ask in plain English. The agent picks tools, calls them, and shows live JSON.

See it

Test MCP with AI agents

Run your server across 40+ models — Claude, GPT, Gemini, Grok — side by side.

See compare mode

Configure & save an agent

Pin a model + servers + system prompt as a reusable agent. Reopen in one click.

See agent setup
Hobby API

Export as API

Call your saved agent from your own code — POST to /api/v1/agents/<id>/chat.

See the API
Chat

Chat with your MCP server in plain English

Connect any MCP server — remote HTTP, SSE, or Streamable HTTP — pick a model, and start asking questions. The agent discovers your tools, decides which to call, and streams every invocation live with full JSON input and output. No glue code, no SDK install.

  • Up to 4 MCP servers per chat — tools merged automatically
  • Live tool-call trace with per-call latency and full JSON
  • Switch model mid-chat to compare reasoning quality
Open Agent Studio
Chat with your MCP server: model picker, server connector, and live tool calls streaming as the AI agent answers.
Compare Mode: same MCP prompt running across multiple AI models in parallel with live tool calls and winner ribbons.
Test MCP with AI

Test your MCP server across 40+ AI models — in one run

Send the same prompt to up to 4 models in parallel: Claude, GPT-5, Gemini, DeepSeek, Grok and more. See every tool call live, compare outputs side-by-side, and let automatic winner ribbons surface the fastest, cheapest, and most accurate model for your tools.

  • Up to 4 models per run — Claude, GPT, Gemini, DeepSeek, Grok, Qwen…
  • Auto-computed winners: Fastest, Cheapest, Best Accuracy
  • See exactly which tools each model called and why
Try Compare Mode
Save & Reuse AI Agents

Configure once, reopen forever

Pin a model, MCP servers, and a system prompt as a named agent — “GitHub Code Reviewer”, “Stripe Refund Assistant”, “Linear Standup”. Reopen any saved agent in one click from your sidebar, or start from one of dozens of pre-built templates and customize from there.

  • Pre-built templates for GitHub, Stripe, Linear, Slack, Notion + more
  • One-click reopen — model, servers and prompt restored exactly
  • Per-agent system prompts you can iterate on
Browse agent templates
Save as Agent
GitHub Code Reviewer
Claude Sonnet 4.6
🐙 GitHub📋 Linear
You are a thorough code reviewer. For each PR, summarise risk, flag missing tests, and link related Linear issues…
Saved · reopen anytime

My agents

🐙 GitHub Code Reviewer💳 Stripe Refund Assistant📋 Linear Standup📊 Analytics Q&A
Terminal · curlHobby API

POST endpoint

https://mcpplaygroundonline.com/api/v1/agents/<AGENT_ID>/chat
$ curl https://mcpplaygroundonline.com/api/v1/agents/agt_8f3d/chat \
   -H "Authorization: Bearer mcpp_••••••••••" \
   -H "Content-Type: application/json" \
   -d '{ "message": "Top 5 open PRs?" }'

# → streamed response with tool calls + final answer
{ "role": "assistant",
  "text": "Found 5 open PRs. Top: #312 Auth timeout (P0)…",
  "tool_calls": [{ "name": "list_pull_requests" }]
}
Export AI Agent with MCPs

Export your agent as a callable API

Every saved agent gets a stable endpoint. Generate an API key in settings, then call /api/v1/agents/<id>/chat from your backend, scripts, or worker. Same model, same MCP servers, same system prompt — now driven by your code.

  • Bearer-token auth with per-org keys you can rotate anytime
  • Locked-in system prompt — your guardrails are server-side and can’t be overridden through API input
  • Pre-configured safety: pagination caps, tool-failure handling, and per-server tool isolation on every call
  • Reuses the agent you tested in Studio — no drift between dev and prod
  • Streamed responses with tool-call traces in the payload
Get an API key

Who is this for?

Pick your starting point

Add one or more MCP servers, pick any AI model, and ask in plain English. Inspect every live tool call, compare models side-by-side, and find the best quality-to-cost fit for your workflow.

Test & Compare with AI

For MCP developers

Connect your server and run it against Claude, GPT, Gemini & 40+ models side-by-side. See every tool call live before you ship.

Build a Reusable Agent

For teams & power users

Start from a Stripe, GitHub, or Linear template. Pick your model, save the config, and use it every day — no re-setup.

Sign in with email

Free credits on sign-up · No credit card · Terms

Or browse agent templates — no sign-in needed.

Templates

Start from a pre-built agent

Ready-to-run templates for the tools you already use — click one, add your access token, and start chatting in under a minute. No sign-in needed to browse.

Knowledge base

Notion Knowledge Agent

Chat with your Notion workspace — search pages, summarise docs, create new entries and keep your knowledge base tidy.

Single serverView →
Most popular

GitHub Code Review

Review PRs, inspect diffs, summarise commits and suggest improvements across any repository.

Single serverView →
Error monitoring

Sentry Error Triage

Triage Sentry errors in natural language — inspect stack traces, group issues and spot the commits that caused them.

Single serverView →
Productivity

Linear Project Manager

Create issues, triage backlogs, track progress and summarise sprint status directly from chat.

Single serverView →
CRM

HubSpot CRM Assistant

Query contacts, deals and companies in HubSpot — update records, spot pipeline risks and draft personalised outreach.

Single serverView →
Commerce

Shopify Storefront

Browse a Shopify store’s catalog, search products and build a cart from chat — powered by the store’s own hosted Storefront MCP.

Single serverView →
E-commerce

Amazon Ads Operator

Query Sponsored Products, Brands and Display campaigns in plain English — bid adjustments, keyword analysis and performance reports without Seller Central tab-switching.

Single serverView →
3 servers

Paid Media ROAS · Meta, Amazon & Stripe

Meta Ads, Amazon Ads and Stripe in one chat — see true blended ROAS across Facebook, Instagram and Amazon with live revenue attribution.

3 serversView →
Automation

Zapier Automator

Trigger automations, list Zaps and manage workflows across 6,000+ connected apps.

Single serverView →
3 servers

Full-Stack Shipper · GitHub, Vercel & Sentry

GitHub PRs, Vercel deploys and Sentry errors in one chat — the complete Next.js ship-and-debug loop.

3 serversView →
Browse all templates

Hosted MCPs

Need a hosted MCP server URL to test?

To test a hosted MCP server, paste its URL (HTTP, SSE, or Streamable HTTP) into MCP Agent Studio — a no-install browser MCP client — or deploy your own sandboxed hosted server with a live URL in under a minute.

Official hosted MCP server endpoints you can paste straight into MCP Agent Studio.
ServiceOfficial hosted MCP URLAuth
GitHubhttps://api.githubcopilot.com/mcp/PAT / OAuth
Notionhttps://mcp.notion.com/mcpOAuth
Stripehttps://mcp.stripe.com/API key
Supabasehttps://mcp.supabase.com/mcpOAuth / PAT
Linearhttps://mcp.linear.app/sseOAuth
Sentryhttps://mcp.sentry.dev/mcpOAuth
Figmahttps://mcp.figma.com/mcpOAuth
Neonhttps://mcp.neon.tech/sseOAuth
PostgreSQLuser-provided URL — or deploy a hosted sandboxBearer token

Or deploy your own sandboxed MCP server

database

MongoDB

Full read/write access to a MongoDB cluster.

SandboxedDeploy →
database

Postgres

Read-only SQL queries against any reachable database.

SandboxedDeploy →
saas

GitHub

Browse repos, read code, manage issues and pull requests.

SandboxedDeploy →
saas

Slack

Post messages, read channels, manage threads in a workspace.

SandboxedDeploy →
browser

Playwright

Drive a real Chrome browser — navigate, click, fill, screenshot.

SandboxedDeploy →
saas

Stripe

Read and manage customers, charges, subscriptions, invoices.

SandboxedDeploy →
saas

Notion

Read and write pages, databases, and blocks in a Notion workspace.

SandboxedDeploy →
search

Wikipedia

Search and read Wikipedia articles — zero-config.

SandboxedDeploy →
Browse all hosted MCP servers

40+

AI Models

Pre-built

Agent Templates

4

MCP Servers per Chat

Free

To Get Started

Model choice

Which AI model is best for MCP tool calling?

There is no single best model for MCP tool calling — it depends on your server's tools. MCP Agent Studio runs the same prompt on up to 4 models in parallel and scores accuracy, speed, and cost automatically.

Claude Sonnet 4.6 and Claude Opus 4.6 are the strongest default tool-callers in our catalog; GPT-5.4 and GPT-5.5 lead on structured-output tasks; Gemini 3.1 Pro handles long tool lists well; DeepSeek V4, GLM 5.1, and Nemotron 3 are the budget picks that still call tools reliably. Instead of guessing, paste your MCP server URL, pick 4 models, and let Compare Mode's winner ribbons (Fastest / Cheapest / Best Accuracy) decide. Works with Claude, GPT-5, Gemini, Grok, Qwen, Kimi, MiniMax, Mistral — 40+ models total.

Run any MCP server with GPT-5, Claude, or Gemini — no SDK, no install

Claude Desktop only speaks to Claude and ChatGPT only speaks to OpenAI models. MCP Agent Studio is a browser MCP client that connects the same server to any model — switch mid-conversation without reconfiguring.

Server × model

Which AI model should I use with the GitHub, Notion, or Postgres MCP server?

Every popular MCP server — GitHub, Notion, Postgres, Stripe, Supabase, Linear, Slack — works with every major AI model in MCP Agent Studio. Pick a server below for the recommended model, its official hosted URL, and a ready-made agent template.

MCP server × AI model quick-start matrix — tested in MCP Agent Studio's Compare Mode. “Recommended” = best accuracy on the server's typical tasks; “Budget” = cheapest model that still calls tools reliably.
MCP serverRecommended modelBudget pickOfficial hosted URLStart from
GitHub MCP serverClaude Sonnet 4.6GPT-5.4 minihttps://api.githubcopilot.com/mcp/GitHub Code Review
Notion MCP serverGPT-5.4Gemini 3 Flashhttps://mcp.notion.com/mcpNotion Knowledge Agent
PostgreSQL MCP serverClaude Sonnet 4.6DeepSeek V4 Flashuser-provided URL — or hosted sandboxData Analyst
Stripe MCP serverClaude Opus 4.6GLM 5.1https://mcp.stripe.com/Stripe Billing Assistant
Supabase MCP serverClaude Sonnet 4.6Gemini 3 Flashhttps://mcp.supabase.com/mcpSupabase Builder
Linear MCP serverGPT-5.4GPT-5.4 nanohttps://mcp.linear.app/sseLinear Project Manager
Slack MCP serverGemini 3.1 ProQwen 3.7 Plushttps://mcp.slack.com/mcpSlack Team Assistant
Sentry MCP serverClaude Sonnet 4.6DeepSeek V4 Prohttps://mcp.sentry.dev/mcpSentry Error Triage
Figma MCP serverGemini 3.1 ProGemini 3 Flashhttps://mcp.figma.com/mcp—
Neon MCP serverClaude Sonnet 4.6Nemotron 3 Super 120Bhttps://mcp.neon.tech/sseNeon Data Ops

Also works with Grok, Kimi, MiniMax, Mistral, GLM and every other model in the catalog — and with the Cloudflare, Vercel, HubSpot, Jira & Confluence, Asana, Monday.com, MongoDB, Playwright, Salesforce, Zapier, Docker MCP servers.

Related free toolsMCP server tester · online MCP inspector · MCP security scanner · MCP token counter

FAQ

Frequently asked questions

Everything you might want to know about using AI agents with MCP servers.

What is MCP Agent Studio?

MCP Agent Studio is a browser-based AI agent that connects to any Model Context Protocol (MCP) server and lets you query your data in plain English using 40+ AI models including Claude, GPT-5, Gemini, DeepSeek, Mistral, Grok, Qwen, NVIDIA Nemotron, and more. Sign up for free — no credit card required.

Do I need to write code to use MCP Agent Studio?

No. MCP Agent Studio is a no-code interface. You paste your MCP server URL, pick an AI model, and start chatting. The agent discovers the available tools automatically and calls them for you as needed.

Which AI models can I use with MCP Agent Studio?

You can use 40+ models from every major provider, including Claude Opus 4.6, Claude Sonnet 4.6, GPT-5.4, GPT-5.5, Gemini 3.1 Pro, DeepSeek V4, Mistral Small 2603, Grok 4.5, Qwen 3, NVIDIA Nemotron Super 120B, Google Gemma 4, Z.AI GLM 4.5 Air, and more. You can switch models mid-conversation to compare responses on the same task.

What is an MCP server and do I need one?

An MCP server is a lightweight service that exposes tools (functions, database queries, API calls) over the Model Context Protocol so any AI agent can use them. Popular public MCP servers include GitHub, Supabase, Notion, Slack, Stripe, Postgres, Linear, and Playwright. You can also connect your own custom MCP server over HTTP, SSE, or Streamable HTTP.

How does pricing work?

MCP Agent Studio is free to sign up — no credit card required. Running AI models costs credits, which you can purchase in packs from your Settings page. New accounts may receive a small starter credit to try the product. Credit costs vary by model; check the Settings › Credits page for current pricing.

Can I query my own database with MCP Agent Studio?

Yes. Connect the official Postgres or Supabase MCP server (or any MCP server that wraps your database) and ask questions in plain English: "Show me my top 10 customers by revenue this month." The AI generates the SQL, runs it through the MCP server, and returns the results with a natural-language summary.

How is this different from ChatGPT or Claude Desktop?

Claude Desktop only works with Claude. ChatGPT only works with OpenAI models. MCP Agent Studio works with any model from any provider — swap between Claude, GPT, Gemini, DeepSeek, and Grok on the same MCP server without re-configuring anything. It also runs entirely in your browser with no local installation.

Can I run an MCP server with GPT-5?

Yes. MCP Agent Studio connects any MCP server (HTTP, SSE, or Streamable HTTP) to GPT-5.4, GPT-5.5, GPT-5 mini, and the rest of the GPT-5 family. Paste your server URL, pick the model, and every tool call streams live — no OpenAI SDK or glue code required.

Which AI model is best for MCP tool calling?

It depends on your server. Claude Sonnet 4.6 and Opus 4.6 are the strongest default tool-callers; GPT-5.4 leads on structured outputs; Gemini 3.1 Pro handles long tool lists; DeepSeek V4 and GLM 5.1 are the budget picks. Compare Mode in MCP Agent Studio scores all of them on your actual server.

Is there a hosted MCP server URL I can test with?

Yes. Official hosted endpoints include Notion (https://mcp.notion.com/mcp), Stripe (https://mcp.stripe.com/), GitHub (https://api.githubcopilot.com/mcp/), Linear (https://mcp.linear.app/sse), and Sentry (https://mcp.sentry.dev/mcp). Paste any of them into MCP Agent Studio, or deploy your own sandboxed hosted MCP server with a live URL in under a minute.

How do I connect to a remote or hosted MCP server?

Paste the server URL into MCP Agent Studio, add an auth token if the server requires one, and click connect — the agent lists the tools automatically. Remote HTTP, SSE, and Streamable HTTP transports are all supported, entirely in the browser with nothing to install.

Can I use the GitHub MCP server with Claude?

Yes — and with GPT-5, Gemini, Grok, or any of 40+ models. MCP Agent Studio connects the official GitHub MCP server (https://api.githubcopilot.com/mcp/) to whichever model you pick, so you can review PRs or triage issues with Claude and re-run the same prompt on GPT-5 to compare.

Does MCP work with Gemini?

Yes. The Model Context Protocol is model-agnostic, and MCP Agent Studio acts as a Gemini MCP client: connect any server and run it with Gemini 3.1 Pro or Gemini 3 Flash. Gemini Flash is one of the cheapest models that still calls MCP tools reliably.

Is my data secure?

Your MCP server credentials stay in your browser session and are never stored on our servers. Conversations are scoped to your Supabase-authenticated organization. We never train on your data, and you can delete your history at any time from Settings.

Can I compare multiple AI models on the same task?

Yes — this is one of the core reasons people use MCP Agent Studio. Run the same prompt against Claude, GPT-5, and Gemini side by side, inspect every tool call in the live trace panel, and find the cheapest model that still gives great results. Teams routinely cut inference costs 5–10× this way.

What MCP servers work with Agent Studio?

Any MCP server that speaks HTTP, SSE, or Streamable HTTP. That includes official servers for GitHub, Supabase, Postgres, Notion, Slack, Stripe, Linear, Playwright, Filesystem, Memory, Sequential Thinking, and hundreds of community servers. You can also point it at your own custom server deployed on Cloudflare Workers, Vercel, Fly, or anywhere else.

Do you have pre-built templates I can use?

Yes. MCP Agent Studio ships with ready-to-run templates covering Notion, GitHub, Amazon Ads, Jira, Confluence, Sentry, HubSpot, Stripe, Linear, Slack, Supabase, Vercel, Shopify, Cloudflare, Zapier, and more — including multi-server combos like GitHub + Vercel + Sentry and Amazon Ads + Stripe for blended ROAS. Pick a template, add your access token, and start chatting in seconds. Browse all templates at /templates.

Can I save an agent configuration and reuse it?

Yes. Once you set up a model, one or more MCP servers, and a system prompt, you can save the whole configuration as a named agent with one click. Open it again any time from "My Agents" — your servers, tools and model selection are all restored. Agents can be private or shared with your team.

Can I export my agent as code?

Export is coming soon. Save your agent and open Export in the studio toolbar to preview a Next.js App Router layout or a Python starter (MCP wiring, prompts, .env.example, deploy targets). Downloadable ZIP packs and one-click deploy will ship in a follow-up release; until then you can explore the preview and recreate wiring manually if you self-host.

Get started in seconds

Start chatting with your data — free

No credit card required. Connect any MCP server and start querying it with Claude, GPT-5, or Gemini in under a minute.

Sign in with email

Free credits on sign-up · No credit card · Terms

MCP Agent Studio — Run MCP Servers with Claude, GPT-5 & Gemini | MCP Playground