API v1

API Documentation

Generate system architecture diagrams programmatically. One endpoint, JSON in, JSON out.

Endpoints

POSThttps://codelit.io/api/generate

Generate a system architecture from a text description. Returns JSON with nodes, edges, and metadata.

POSThttps://codelit.io/api/github/analyze

Analyze a GitHub repository. Returns detected files, tech patterns, directory structure, and source files for architecture generation.

Body: { owner, repo, branch } · Header: x-github-token (optional for public repos)

Request Body

Field	Type	Required	Description
prompt	string	Yes	System description (3-500 chars)
apiKey	string	No	Your OpenRouter API key (optional)

Response

{
  "architecture": {
    "title": "Netflix Video Streaming Architecture",
    "description": "Video streaming with CDN...",
    "nodes": [
      { "id": "cdn", "label": "CDN", "type": "cdn", "description": "..." },
      { "id": "api", "label": "API Gateway", "type": "backend", "description": "..." }
    ],
    "edges": [
      { "from": "cdn", "to": "api", "label": "HTTPS", "dataFlow": "high" }
    ]
  },
  "model": "nvidia/nemotron-3-super-120b-a12b:free",
  "usage": { "prompt_tokens": 300, "completion_tokens": 800 }
}

Node Types

frontendbackenddatabasequeuecachecdnexternalservice

Examples

cURL

curl -X POST https://codelit.io/api/generate \
  -H "Content-Type: application/json" \
  -d '{"prompt": "Netflix video streaming architecture"}'

JavaScript

const res = await fetch("https://codelit.io/api/generate", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    prompt: "Uber ride matching at scale"
  }),
});
const { architecture } = await res.json();
console.log(architecture.title, architecture.nodes.length, "nodes");

Python

import requests

res = requests.post("https://codelit.io/api/generate", json={
    "prompt": "Spotify music streaming platform"
})
arch = res.json()["architecture"]
print(f"{arch['title']} — {len(arch['nodes'])} nodes")

GitHub Analyze (cURL)

curl -X POST https://codelit.io/api/github/analyze \
  -H "Content-Type: application/json" \
  -d '{"owner": "vercel", "repo": "next.js", "branch": "canary"}'

GitHub Analyze (JS)

const res = await fetch("https://codelit.io/api/github/analyze", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    owner: "vercel", repo: "next.js", branch: "canary"
  }),
});
const { files, patterns, totalFiles } = await res.json();
console.log(`${totalFiles} files, detected: ${patterns.join(", ")}`);

Rate Limits

The API uses free AI models with automatic fallback. Rate limits depend on upstream model providers.

For higher limits, pass your own apiKey (OpenRouter key).

Playground

Try it live

Try it now

No API key required. Just send a POST request.

Try the App View Pricing

{ }

Endpoints

POSThttps://codelit.io/api/generate

Generate a system architecture from a text description. Returns JSON with nodes, edges, and metadata.

POSThttps://codelit.io/api/github/analyze

Analyze a GitHub repository. Returns detected files, tech patterns, directory structure, and source files for architecture generation.

Body: { owner, repo, branch } · Header: x-github-token (optional for public repos)

Field

Type

Required

Description

prompt

string

Yes

System description (3-500 chars)

apiKey

string

Your OpenRouter API key (optional)

Response

{ "architecture": { "title": "Netflix Video Streaming Architecture", "description": "Video streaming with CDN...", "nodes": [ { "id": "cdn", "label": "CDN", "type": "cdn", "description": "..." }, { "id": "api", "label": "API Gateway", "type": "backend", "description": "..." } ], "edges": [ { "from": "cdn", "to": "api", "label": "HTTPS", "dataFlow": "high" } ] }, "model": "nvidia/nemotron-3-super-120b-a12b:free", "usage": { "prompt_tokens": 300, "completion_tokens": 800 } }

Examples

cURL

curl -X POST https://codelit.io/api/generate \
  -H "Content-Type: application/json" \
  -d '{"prompt": "Netflix video streaming architecture"}'

JavaScript

const res = await fetch("https://codelit.io/api/generate", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    prompt: "Uber ride matching at scale"
  }),
});
const { architecture } = await res.json();
console.log(architecture.title, architecture.nodes.length, "nodes");

Python

import requests

res = requests.post("https://codelit.io/api/generate", json={
    "prompt": "Spotify music streaming platform"
})
arch = res.json()["architecture"]
print(f"{arch['title']} — {len(arch['nodes'])} nodes")

GitHub Analyze (cURL)

curl -X POST https://codelit.io/api/github/analyze \
  -H "Content-Type: application/json" \
  -d '{"owner": "vercel", "repo": "next.js", "branch": "canary"}'

GitHub Analyze (JS)

const res = await fetch("https://codelit.io/api/github/analyze", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    owner: "vercel", repo: "next.js", branch: "canary"
  }),
});
const { files, patterns, totalFiles } = await res.json();
console.log(`${totalFiles} files, detected: ${patterns.join(", ")}`);