extract-openapi-from-code
extract-openapi-from-code
Extract an OpenAPI specification from an existing API codebase. Covers eight major frameworks across Python, Java, JavaScript/TypeScript, Ruby, and PHP.
Content Guides
| Framework | Language | Guide |
|---|---|---|
| FastAPI | Python | content/frameworks/fastapi.md |
| Flask | Python | content/frameworks/flask.md |
| Django REST Framework | Python | content/frameworks/django.md |
| Spring Boot | Java | content/frameworks/spring-boot.md |
| NestJS | TypeScript | content/frameworks/nestjs.md |
| Hono | TypeScript | content/frameworks/hono.md |
| Rails | Ruby | content/frameworks/rails.md |
| Laravel | PHP | content/frameworks/laravel.md |
Each guide provides detailed setup, schema definition, Speakeasy extensions, authentication, and troubleshooting for that framework.
When to Use
- User has an existing API and wants to generate an OpenAPI spec from it
- User wants to create an SDK from code that has no OpenAPI spec yet
- User mentions a specific framework (FastAPI, Flask, Django, Spring Boot, NestJS, Hono, Rails, Laravel)
- User says: "extract OpenAPI", "code first", "generate spec from code", "existing API code"
Inputs
| Input | Required | Description |
|---|---|---|
| Framework | Yes | The API framework in use (see Decision Framework) |
| Project path | Yes | Root directory of the API project |
| Output path | No | Where to write the spec (default: openapi.json or openapi.yaml) |
| Target language | No | SDK target language, if generating an SDK after extraction |
Outputs
| Output | Description |
|---|---|
| OpenAPI spec | A JSON or YAML file describing the API |
| Validation report | Lint results from speakeasy lint |
Prerequisites
- The API project must be buildable and its dependencies installed
- For runtime extraction (FastAPI, Spring Boot, NestJS, Hono), the app must be importable or startable
speakeasyCLI installed for post-extraction validation and SDK generation
Decision Framework
Use this tree to determine the extraction method:
| Framework | Language | Method | Requires Running Server? |
|---|---|---|---|
| FastAPI | Python | Built-in export | No |
| Flask (flask-smorest) | Python | CLI command | No |
| Django REST Framework | Python | drf-spectacular CLI | No |
| Spring Boot (springdoc) | Java | HTTP endpoint | Yes |
| NestJS | TypeScript | HTTP endpoint or script | Yes |
| Hono (zod-openapi) | TypeScript | Programmatic export | No |
| Rails (rswag) | Ruby | Rake task | No |
| Laravel (l5-swagger) | PHP | Artisan command | No |
Command
Choose the command matching your framework below. After extraction, always validate with speakeasy lint.
Python: FastAPI
FastAPI generates an OpenAPI schema at runtime. Export it without starting the server:
python -c "import json; from myapp import app; print(json.dumps(app.openapi()))" > openapi.json
Replace myapp with the module containing your FastAPI app instance. If the app uses a factory pattern:
python -c "import json; from myapp import create_app; app = create_app(); print(json.dumps(app.openapi()))" > openapi.json
You can also start the server and fetch from http://localhost:8000/openapi.json.
Python: Flask (flask-smorest)
Requires flask-smorest or apispec:
flask openapi write openapi.json
If using apispec directly, export programmatically:
import json
from myapp import create_app, spec
app = create_app()
with app.app_context():
print(json.dumps(spec.to_dict()))
Python: Django REST Framework
Requires drf-spectacular:
python manage.py spectacular --file openapi.yaml
For JSON output:
python manage.py spectacular --format openapi-json --file openapi.json
Java: Spring Boot
Requires springdoc-openapi. Start the application, then fetch the spec:
# Start the app (background)
./mvnw spring-boot:run &
# Wait for startup
sleep 15
# Fetch the spec
curl http://localhost:8080/v3/api-docs -o openapi.json
# For YAML format
curl http://localhost:8080/v3/api-docs.yaml -o openapi.yaml
# Stop the app
kill %1
If the server runs on a different port or context path, adjust the URL accordingly.
TypeScript: NestJS
Requires @nestjs/swagger. Start the application, then fetch:
# Start the app (background)
npm run start &
sleep 10
# Fetch the spec (default path with SwaggerModule)
curl http://localhost:3000/api-json -o openapi.json
# Stop the app
kill %1
Alternatively, create a script to export without running the server:
// scripts/export-openapi.ts
import { NestFactory } from '@nestjs/core';
import { SwaggerModule, DocumentBuilder } from '@nestjs/swagger';
import { AppModule } from '../src/app.module';
import * as fs from 'fs';
async function bootstrap() {
const app = await NestFactory.create(AppModule, { logger: false });
const config = new DocumentBuilder().setTitle('API').build();
const doc = SwaggerModule.createDocument(app, config);
fs.writeFileSync('openapi.json', JSON.stringify(doc, null, 2));
await app.close();
}
bootstrap();
TypeScript: Hono (zod-openapi)
Requires @hono/zod-openapi. Export the schema programmatically:
// scripts/export-openapi.ts
import { app } from '../src/app';
import * as fs from 'fs';
const doc = app.doc('/doc', {
openapi: '3.1.0',
info: { title: 'API', version: '1.0.0' },
});
fs.writeFileSync('openapi.json', JSON.stringify(doc, null, 2));
Run with:
npx tsx scripts/export-openapi.ts
Ruby: Rails (rswag)
Requires rswag:
rails rswag:specs:swaggerize
The spec is written to swagger/v1/swagger.yaml by default (configurable in config/initializers/rswag_api.rb).
PHP: Laravel (l5-swagger)
Requires l5-swagger:
php artisan l5-swagger:generate
The spec is written to storage/api-docs/api-docs.json by default.
Post-Extraction Steps
After extracting the spec, always run these steps:
1. Validate the spec
speakeasy lint openapi --non-interactive -s openapi.json
2. Fix issues with overlays (if needed)
If validation reveals issues, use an overlay rather than modifying the extracted spec directly:
speakeasy overlay apply -s openapi.json -o fixes.yaml
To fix validation errors, create an OpenAPI overlay file and apply it with speakeasy overlay apply -s <spec> -o <overlay>.
3. Generate an SDK
speakeasy quickstart --skip-interactive --output console \
-s openapi.json \
-t <target> \
-n <name> \
-p <package>
Run speakeasy quickstart -s <spec> -t <language> to initialize a new SDK project.
Example
Full workflow for a FastAPI project:
# 1. Extract the OpenAPI spec
cd /path/to/my-fastapi-project
python -c "import json; from main import app; print(json.dumps(app.openapi()))" > openapi.json
# 2. Validate
speakeasy lint openapi --non-interactive -s openapi.json
# 3. Generate a TypeScript SDK
speakeasy quickstart --skip-interactive --output console \
-s openapi.json \
-t typescript \
-n "MyApiSDK" \
-p "my-api-sdk"
Adding Speakeasy Extensions
After extracting a spec, add Speakeasy-specific extensions for better SDK output. These can be added in framework config or via overlay.
FastAPI: Add Extensions via openapi_extra
@app.get(
"/items",
openapi_extra={
"x-speakeasy-retries": {
"strategy": "backoff",
"backoff": {"initialInterval": 500, "maxInterval": 60000, "exponent": 1.5},
"statusCodes": ["5XX", "429"]
},
"x-speakeasy-group": "items",
"x-speakeasy-name-override": "list"
}
)
def list_items(): ...
Django: Add Extensions via SPECTACULAR_SETTINGS
# settings.py
SPECTACULAR_SETTINGS = {
# ... other settings
'EXTENSIONS_TO_SCHEMA_FUNCTION': lambda generator, request, public: {
'x-speakeasy-retries': {
'strategy': 'backoff',
'backoff': {'initialInterval': 500, 'maxInterval': 60000, 'exponent': 1.5},
'statusCodes': ['5XX']
}
}
}
Spring Boot: Add Extensions via Custom OperationCustomizer
@Bean
public OperationCustomizer operationCustomizer() {
return (operation, handlerMethod) -> {
operation.addExtension("x-speakeasy-group",
handlerMethod.getBeanType().getSimpleName().replace("Controller", "").toLowerCase());
return operation;
};
}
NestJS: Add Extensions via Decorator Options
@Get()
@ApiOperation({
summary: 'List items',
operationId: 'listItems'
})
@ApiExtension('x-speakeasy-group', 'items')
@ApiExtension('x-speakeasy-name-override', 'list')
listItems() { ... }
Via Overlay (Any Framework)
If you cannot modify framework code, use an overlay:
overlay: 1.0.0
info:
title: Speakeasy Extensions
version: 1.0.0
actions:
- target: $.paths['/items'].get
update:
x-speakeasy-group: items
x-speakeasy-name-override: list
Common Issues After Extraction
| Issue | Symptom | Fix |
|---|---|---|
| Missing operationIds | Lint warning; SDK methods get generic names | Add operationIds via overlay or use speakeasy suggest operation-ids -s openapi.json |
| Missing descriptions | Lint hints; SDK has no documentation | Add descriptions to endpoints and schemas in source code or via overlay |
| Overly permissive schemas | Schemas use additionalProperties: true or lack type constraints |
Tighten schemas in source code; use stricter validation decorators |
| No response schemas | Lint errors; SDK return types are any/object |
Add explicit response models to your framework endpoints |
| Duplicate operationIds | Lint errors; generation fails | Ensure each endpoint has a unique operationId |
| Missing authentication | No security schemes in spec | Add security metadata to your framework config or via overlay |
What NOT to Do
- Do NOT hand-write an OpenAPI spec when the framework can generate one -- always extract first
- Do NOT edit the extracted spec directly -- use overlays for fixes so re-extraction does not lose changes
- Do NOT skip validation -- extracted specs often have issues that block SDK generation
- Do NOT assume the extracted spec is complete -- frameworks may omit auth, error responses, or headers
- Do NOT start the server in production mode for extraction -- use development or test configuration
Troubleshooting
| Error | Cause | Solution |
|---|---|---|
ModuleNotFoundError (Python) |
App dependencies not installed | Run pip install -r requirements.txt or pip install -e . |
| Connection refused (Spring Boot, NestJS) | Server not fully started | Increase sleep time or poll for readiness |
| Empty or minimal spec | Routes not registered at import time | Ensure all route modules are imported; check lazy loading |
| YAML parse error | Extracted file has invalid syntax | Re-extract; check for print statements polluting stdout |
Cannot find module (Node.js) |
Dependencies not installed | Run npm install or yarn install |
No /v3/api-docs endpoint (Spring Boot) |
springdoc not configured | Add springdoc-openapi-starter-webmvc-ui to dependencies |
No /api-json endpoint (NestJS) |
Swagger module not set up | Configure SwaggerModule.setup(app, ...) in main.ts |
Related Skills
manage-openapi-overlays- Add x-speakeasy-* extensions via overlaystart-new-sdk-project- Generate SDK after extraction
More from speakeasy-api/skills
writing-openapi-specs
Reference guide for OpenAPI specification best practices, naming conventions, and expressing complex REST API patterns like polymorphism, enums, file uploads, and server-sent events. Use when writing or improving OpenAPI specs to ensure they follow established conventions and generate quality SDKs.
77speakeasy-context
Speakeasy workflow: run 'agent context' FIRST, do task, run 'agent feedback' LAST. Triggers on speakeasy, SDK, OpenAPI.
51diagnose-generation-failure
Use when SDK generation failed or seeing errors. Triggers on "generation failed", "speakeasy run failed", "SDK build error", "workflow failed", "Step Failed", "why did generation fail
50manage-openapi-overlays
Use when creating, applying, or validating overlay files including x-speakeasy extensions. Covers overlay syntax, JSONPath targeting, retries, pagination, naming, grouping, open enums, global headers, custom security. Triggers on "create overlay", "apply overlay", "overlay file", "x-speakeasy", "add extension", "configure retries", "add pagination", "overlay for retries".
49customize-sdk-hooks
|
48generate-mcp-server
Use when generating an MCP server from an OpenAPI spec with Speakeasy. Triggers on "generate MCP server", "MCP server", "Model Context Protocol", "AI assistant tools", "Claude tools", "speakeasy MCP", "mcp-typescript
48