Kong Insomnia 12: Smarter, Faster, More Accessible API and MCP Development

• We spent the summer investing in our core experiences like Git Sync, SSO/SCIM, and more. Check out our changelog for all the details. We’re ready for you!
• Looking for best practices in Insomnia? Check out our new Youtube playlist.

When you’re building an MCP server that provides tools, resources, or prompts that AI agents can consume, you’ll need to verify that:

• Your server actually works. 
• Your tools behave as expected. 
• Your error handling works. 

In Insomnia 12, we’re bringing the same collaboration and confidence to MCP development that Insomnia already provides for APIs. As agentic AI becomes critical infrastructure, developers building MCP servers face the same challenges API developers solved years ago: how do you quickly test and validate what you’re building without a complicated setup? Insomnia’s new MCP client brings the familiar test-iterate-debug workflow to AI-native development. 

Connect directly to your MCP servers, manually invoke any tool with custom parameters, inspect protocol-level & authentication messages, and see responses. 

Here are three example MCP servers if you just want to try one quickly: 

• DeepWiki: https://mcp.deepwiki.com/mcp
• GitHub Copilot (supports oAuth): https://api.githubcopilot.com/mcp/ 
• Notion (supports oAuth and Dynamic Client Registration): https://mcp.notion.com/mcp

Instead of manually defining routes and responses that could take hours, simply describe what you need in natural language, provide a URL, JSON sample, or OpenAPI spec, and Insomnia 12 automatically generates a complete mock server with optional dynamic responses. For teams with a design-first approach, you can easily jump straight from your OpenAPI spec document to a fully functional mock server. 

Testing against a metered API like Stripe? Developing against an internal service that’s complex to run locally? Need a mock server for a feature that’s still being built? Describe it and you’re testing in seconds. 

What used to require hours of manual set up, like defining endpoints or crafting realistic responses, now happens almost instantaneously with AI. Mock servers can now transform from a “nice to have if you have the time to set them up” into an essential part of a developer’s workflow, allowing you to test faster without manual overhead. 

Check out the following demo videos showing the three ways to autogenerate mock servers.

Writing good commit messages takes time — time that developers would rather spend building. Under pressure or lack of time, commits become rushed afterthoughts, creating a messy git history. A messy git history makes collaborating and debugging later on harder for everyone.

Insomnia 12 can now analyze your diffs and commit history to automatically suggest logical file groupings with clear, descriptive commit messages. Accept them as-is, edit to fit your style, or ignore them entirely. Either way, your git history stays clean and your team stays focused on the code. Your code, your choice.

We think it’s important to be transparent in how we use AI, how you can configure it, and to put you in total control of where your code and data is used or stored.

• AI features can be enabled or disabled per account or per user. Our AI features are helpfully marked with a sparkles icon so you always know when you’re using them.
• You can configure whether to use a cloud provider (Anthropic, OpenAI, Gemini) or use a Local LLM. Local LLMs mean the data never leaves your machine, at the cost of being moderately less performant. With Insomnia, you are always in control of where your code and data go.
• As with any AI-powered functionality, you should carefully review its output before use, and different models might provide different quality results.
• Our AI features are in early access right now, and we’d love your feedback. They're also free for now, but we may choose to adjust how they're available at a later date.

Testing and debugging APIs (and now MCP servers) is collaborative work, and we’re making it easier to bring your team along. Git Sync, essential for version control, cross-machine syncing, and team collaboration, is now available for free for 3 users on the Essentials (formerly Hobby) plan (or stick to Cloud/Local projects and keep as many users as you want). 

Whether you’re a start up developer scaling your API infrastructure or a small team building your first integration, you can now work together with the same collaboration tools larger orgs rely on, without the price tag. 

Want to explore our full enterprise capabilities? Start a free 14-day Enterprise trial instantly to test advanced security, governance & compliance features like SCIM, SSO, and RBAC. When you’re ready to commit and bring on your team, self-serve up to 50 seats in minutes and start collaborating immediately. 

Check out our new pricing page to find the plan that fits your team’s needs and start collaborating immediately. What you see is what you pay, no hidden fees, no a la carte add-ons, just transparent pricing so you know exactly what you’re paying for.

For the full list of features, fixes, and updates, please see the available changelog. If you have feedback or questions about our new or upcoming releases, reach out in our Slack Community.