Skip to main content
AI tools generate better code when they have access to accurate API documentation. Without proper context, you might get:
  • Incorrect endpoint URLs
  • Wrong parameter names
  • Missing authentication headers
Our one-stop API spec provides your tools with up-to-date documentation and best practices for both HTTP and WebSocket APIs.

How to provide context

Choose your AI tool below for specific setup instructions:
1

Include our API spec in your prompts to your Augment Agent

Start your prompts with Browse https://api.lmnt.com/spec to tell your Augment Agent to use its browsing tools to pull in our API spec.
Example prompts:
Browse https://api.lmnt.com/spec and create a rust app that reads the latest headlines in a newscaster style from https://text.npr.org/ using the 'brandon' voice

Browse https://api.lmnt.com/spec and write a Python script to create a new voice

Browse https://api.lmnt.com/spec and generate a React component that uses LMNT's WebSocket API for live speech synthesis
1

Open settings

Go to Cursor → Settings → Indexing & Docs
2

Add our API spec

Click “Add doc”, paste https://api.lmnt.com/spec in the URL field, name it LMNT, and click confirm.
3

Start coding

Type @LMNT in your Cursor chat pull in our API context.
Example prompts:
@LMNT Create an app that synthesizes speech with LMNT and streams the audio in real-time

@LMNT Generate code to create a new voice using the LMNT API

@LMNT Show me how to authenticate with the WebSocket API
Pasting @LMNT directly doesn’t work in Cursor’s chat - you’ll need to type @LMNT manually to properly activate the context.
1

Set memory rule

In your Claude Code terminal, type the # symbol followed by your memory rule:
Claude Code
# When working on projects and code related to LMNT, refer to documentation from https://api.lmnt.com/spec
2

Choose user memory or project memory

You’ll be given an option to save the memory to project memory or user memory. Choose project memory if you want LMNT documentation only in your current working project, or choose user memory if you want to apply to all projects.
3

Start coding

Claude Code will automatically reference the LMNT API documentation when you ask for code generation.
Example prompts:
Generate a serverless function that uses LMNT's API for speech synthesis with proper authentication

Build a web application that integrates LMNT's real-time speech capabilities

Create a Python CLI tool that manages LMNT voices and synthesizes speech from text files
1

Include our API spec in your prompt

Start your prompts with Browse https://api.lmnt.com/spec to tell the model to use its browsing tools to pull in our API spec.
Example prompts:
Browse https://api.lmnt.com/spec and create an app that synthesizes speech with LMNT and streams the audio in real-time

Browse https://api.lmnt.com/spec and write a Python script to create a new voice

Browse https://api.lmnt.com/spec and generate a React component that uses LMNT's WebSocket API for live speech synthesis
Some models don’t support web browsing. Watch to make sure that your model actually fetched the spec so it’s not hallucinating.As a fallback you can copy the content of our API spec into your prompt. But maybe find a better model to use!
1

See if your tool has a way to persist external documentation

Some tools have a way to add a link to https://api.lmnt.com/spec that persists between prompts. Use this if is possible.
2

Otherwise, include our API spec in your prompts

Start your prompts with Browse https://api.lmnt.com/spec to tell your tool to use the API spec for reference.
Example prompts:
Browse https://api.lmnt.com/spec and create a mobile app for text-to-speech with voice selection

Browse https://api.lmnt.com/spec and write a rust program to create a new voice
Some tools don’t support web browsing. Watch to make sure that your tool actually fetched the spec so it’s not hallucinating.As a fallback you can copy the content of our API spec into your prompt. But maybe find a better tool to use!
I