Example Function Catalog¶

Verified status as of March 28, 2026. Runtime note: FastFN resolves dependencies and build steps per function: Python uses requirements.txt, Node uses package.json, PHP installs from composer.json when present, and Rust handlers are built with cargo. Host runtimes/tools are required in fastfn dev --native, while fastfn dev depends on a running Docker daemon. This page is a guided tour of the example functions shipped with this repository.

Important distinction:

Public example routes live at normal paths like /hello and /telegram-ai-reply.
Platform control-plane endpoints live under /_fn/* (health, OpenAPI, config, logs).

Run the examples¶

Recommended (Next.js-style routes + showcase):

bin/fastfn dev examples/functions/next-style

Then try:

GET /showcase
GET /openapi.json
GET /docs

Full catalog (everything under examples/functions/):

bin/fastfn dev examples/functions

Note about paths:

When you run a subfolder (like examples/functions/next-style), its routes live at the root (for example /users, /showcase).
When you run the full catalog (examples/functions), each app folder becomes a namespace:
next-style/* is served under /next-style/*
polyglot-tutorial/* is served under /polyglot-tutorial/*
polyglot-db-demo/* is served under /polyglot-db-demo/*

Repository source layout:

Python: examples/functions/python/<name>/
Node: examples/functions/node/<name>/
PHP: examples/functions/php/<name>/
Rust: examples/functions/rust/<name>/
Next.js-style app: examples/functions/next-style/ (file routes)

Those paths describe where this repository stores its examples. They are not the only supported layout for user projects.

Beginner tour (10 minutes)¶

If you're new to FastFN, start here. You only need one terminal.

Start the demo app:

bin/fastfn dev examples/functions/next-style

Open the UI and docs:
GET /showcase (browser)
GET /docs (Swagger UI for public functions)
GET /openapi.json (raw OpenAPI JSON)
Call a JSON endpoint:

curl -sS 'http://127.0.0.1:8080/hello?name=World'

Example response:

{"hello":"saludos World"}

Inspect what the gateway sends to handlers:

curl -sS 'http://127.0.0.1:8080/request-inspector?key=test' \
  -X POST \
  -H 'x-demo: 1' \
  --data 'hello'

You should see JSON including method, path, query, headers (only a safe subset), and body.

Try non-JSON responses:
HTML:

curl -sS 'http://127.0.0.1:8080/html-demo?name=Web'

CSV:

curl -sS 'http://127.0.0.1:8080/csv-demo?name=Alice'

Binary (PNG):

curl -sS 'http://127.0.0.1:8080/png-demo' --output out.png

How these examples work¶

Every example function is “just code” plus optional local config:

handler.py / handler.js / handler.ts etc: the handler file
optional private helper modules such as core.js, _shared.py, lib.php
fn.config.json (optional): methods/routes/handler name/timeouts/etc
fn.env.json (optional): per-function env vars (secrets supported)

Handlers receive a single event object (method/query/headers/body/context/env/client). Handlers return:

JSON helpers: { status, headers, body }
Binary: { status, headers, is_base64: true, body_base64 }
Edge proxy: { proxy: { path, method, headers, body, timeout_ms } } (FastFN performs the upstream fetch)

`custom-handler-demo` (custom handler name, Python + Node variants)¶

This demo shows that you can pick a handler name other than handler using fn.config.json.

Important distinction:

the runtime still resolves the file from entrypoint, file routes, or canonical entry files
invoke.handler only changes which exported symbol is invoked inside that resolved file
Python also accepts main(req) when handler is absent
Python variant:

bin/fastfn dev examples/functions/python/custom-handler-demo
curl -sS 'http://127.0.0.1:8080/custom-handler-demo?name=World'

Expected response:

{"runtime":"python","handler":"main","hello":"World"}

Node variant:

bin/fastfn dev examples/functions/node/custom-handler-demo
curl -sS 'http://127.0.0.1:8080/custom-handler-demo?name=World'

Expected response:

{"runtime":"node","handler":"main","hello":"World"}

Multi-route demo apps¶

These folders contain multiple routes (an “app”), not just one endpoint.

`next-style` (recommended: Next.js-style file routing)¶

Run:

bin/fastfn dev examples/functions/next-style

What it demonstrates:
Next.js-style routing (index.*, [id].*, [...slug].*, method prefixes)
Private helper imports that stay out of OpenAPI (users/_shared.js, blog/_shared.py, php/_shared.php, rust/_shared.rs)
Polyglot handlers living side-by-side (node/python/php/rust)
A small “showcase” UI to click through the demos
Try:
GET /showcase
GET /users
GET /users/123
GET /blog
GET /rust/version

`polyglot-tutorial` (step-by-step multi-runtime pipeline)¶

Run (namespaced under the catalog root):

bin/fastfn dev examples/functions

Try:

curl -sS 'http://127.0.0.1:8080/polyglot-tutorial/step-1'
curl -sS 'http://127.0.0.1:8080/polyglot-tutorial/step-2?name=Ada'
curl -sS 'http://127.0.0.1:8080/polyglot-tutorial/step-3?name=Ada'
curl -sS 'http://127.0.0.1:8080/polyglot-tutorial/step-4'
curl -sS 'http://127.0.0.1:8080/polyglot-tutorial/step-5?name=Ada'

What to look for:

Each step is a different runtime (node -> python -> php -> rust -> node)
Step 5 performs internal HTTP calls to earlier steps and returns a combined flow

`polyglot-db-demo` (shared SQLite across runtimes)¶

Run it directly (routes at the root):

bin/fastfn dev examples/functions/polyglot-db-demo

Try:

curl -sS 'http://127.0.0.1:8080/items'
curl -sS -X POST 'http://127.0.0.1:8080/items' -H 'content-type: application/json' --data '{"name":"first item"}'
curl -sS 'http://127.0.0.1:8080/items'
curl -sS -X PUT 'http://127.0.0.1:8080/items/1' -H 'content-type: application/json' --data '{"name":"updated item"}'
curl -sS -X DELETE 'http://127.0.0.1:8080/items/1'

What to look for:

A single SQLite file is shared across node/python/php/rust handlers
Some internal helper routes are intentionally not public (they require an internal-call header)

`ip-intel` (file routes + optional deps + deterministic mock mode)¶

Run (namespaced under the catalog root):

bin/fastfn dev examples/functions

Try without external network calls:

curl -sS 'http://127.0.0.1:8080/ip-intel/maxmind?ip=8.8.8.8&mock=1'
curl -sS 'http://127.0.0.1:8080/ip-intel/remote?ip=8.8.8.8&mock=1'

Python runtime¶

`cron-tick` (scheduler demo)¶

Route: /cron-tick
Methods: GET
Goal: simple counter you can increment via schedule
Source: examples/functions/python/cron-tick/handler.py

Read current count:

curl -sS 'http://127.0.0.1:8080/cron-tick?action=read'

Example response:

{"function":"cron-tick","action":"read","count":0}

Enable schedule (every 1s) via Console API:

curl -sS 'http://127.0.0.1:8080/_fn/function-config?runtime=python&name=cron-tick' \
  -X PUT -H 'Content-Type: application/json' \
  --data '{"schedule":{"enabled":true,"every_seconds":1,"method":"GET","query":{"action":"inc"},"headers":{},"body":"","context":{}}}'
curl -sS -X POST 'http://127.0.0.1:8080/_fn/reload'

Observe scheduler state:

curl -sS 'http://127.0.0.1:8080/_fn/schedules'

`utc-time` (cron + timezone demo)¶

Route: /utc-time
Methods: GET
Goal: show UTC/local timestamps + scheduler trigger context
Schedule: daily at 09:00 in UTC (cron)
Source: examples/functions/python/utc-time/handler.py

Call it:

curl -sS 'http://127.0.0.1:8080/utc-time'

What to look for:

now_utc vs now_local
trigger (when invoked by the scheduler)

`offset-time` (cron + timezone demo)¶

Route: /offset-time
Methods: GET
Goal: same as utc-time, but scheduled using a fixed offset timezone
Schedule: daily at 09:00 in -05:00 (cron)
Source: examples/functions/python/offset-time/handler.py

Call it:

curl -sS 'http://127.0.0.1:8080/offset-time'

Tip: compare next values via /_fn/schedules, or from the browser devtools:

fetch('/_fn/schedules').then((r) => r.json()).then(console.log)

`tools-loop` (tools loop demo, inspired by agent loops)¶

Route: /tools-loop
Methods: GET, POST
Goal: minimal "agent loop" style planner/executor for testing tools (no API keys).
Default behavior: dry_run=true
Source: examples/functions/python/tools-loop/handler.py

Dry run (plan only):

curl -sS 'http://127.0.0.1:8080/tools-loop?text=quiero%20mi%20ip%20y%20clima&dry_run=true'

What to look for:

plan: selected tools
results: in dry_run=true they are placeholders

Execute tools:

curl -sS 'http://127.0.0.1:8080/tools-loop?tool=ip_lookup,weather&city=Buenos%20Aires&dry_run=false'

Execute tools (offline mock):

curl -sS 'http://127.0.0.1:8080/tools-loop?tool=ip_lookup,weather&city=Buenos%20Aires&dry_run=false&mock=true'

`telegram-ai-reply-py` (Telegram AI bot, Python)¶

Route: /telegram-ai-reply-py
Methods: POST
Goal: Telegram webhook -> OpenAI -> Telegram reply (Python)
Env (secrets): TELEGRAM_BOT_TOKEN, OPENAI_API_KEY
Source: examples/functions/python/telegram-ai-reply-py/handler.py

Example (webhook POST):

curl -sS 'http://127.0.0.1:8080/telegram-ai-reply-py' \
  -X POST \
  -H 'Content-Type: application/json' \
  -d '{"message":{"chat":{"id":123},"text":"Hola"}}'

`hello`¶

Route: /hello
Methods: GET
Query: optional name
Source: examples/functions/python/hello/handler.py

Example:

curl -sS 'http://127.0.0.1:8080/hello?name=World'

Typical response:

{"hello":"saludos World"}

`risk-score`¶

Route: /risk-score
Methods: GET, POST
Inputs:
query.email
fallback header x-user-email
Source: examples/functions/python/risk-score/handler.py

GET example:

curl -sS 'http://127.0.0.1:8080/risk-score?email=user@example.com'

POST example:

curl -sS -X POST 'http://127.0.0.1:8080/risk-score' \
  -H 'x-user-email: user@example.com' \
  -H 'Content-Type: application/json' \
  -d '{}'

Typical response:

{"runtime":"python","function":"risk-score","score":60,"risk":"high","signals":{"email":true,"ip":"172.19.0.1"}}

`slow`¶

Route: /slow
Methods: GET
Query: sleep_ms
Source: examples/functions/python/slow/handler.py

Example:

curl -sS 'http://127.0.0.1:8080/slow?sleep_ms=100'

Example response:

{"runtime":"python","function":"slow","slept_ms":100}

`html-demo`¶

Route: /html-demo
Methods: GET
Content-Type: text/html; charset=utf-8
Source: examples/functions/python/html-demo/handler.py

Example:

curl -sS 'http://127.0.0.1:8080/html-demo?name=Web'

Example response:

<html><body><h1>Hello Web</h1><p>html-demo</p></body></html>

`csv-demo`¶

Route: /csv-demo
Methods: GET
Content-Type: text/csv; charset=utf-8
Source: examples/functions/python/csv-demo/handler.py

Example:

curl -sS 'http://127.0.0.1:8080/csv-demo?name=Alice'

Example response:

id,name,runtime
1,Alice,python

`png-demo`¶

Route: /png-demo
Methods: GET
Content-Type: image/png
Source: examples/functions/python/png-demo/handler.py

Example:

curl -sS 'http://127.0.0.1:8080/png-demo' --output out.png

`lambda-echo`¶

Route: /lambda-echo
Methods: GET
Source: examples/functions/python/lambda-echo/handler.py

Example:

curl -sS 'http://127.0.0.1:8080/lambda-echo?name=Lambda'

Example response:

{"hello":"Lambda"}

`custom-echo`¶

Route: /custom-echo
Methods: GET
Query: v
Source: examples/functions/python/custom-echo/handler.py

Example:

curl -sS 'http://127.0.0.1:8080/custom-echo?v=demo'

Example response:

{"value":"demo"}

`requirements-demo`¶

Route: /requirements-demo
Methods: GET
Dependency hints:
requirements.txt
inline #@requirements comment
Source: examples/functions/python/requirements-demo/handler.py

Example:

curl -sS 'http://127.0.0.1:8080/requirements-demo'

Example response:

{"runtime":"python","function":"requirements-demo"}

`qr`¶

Route: /qr
Methods: GET
Query: text or url
Content-Type: image/svg+xml
Source: examples/functions/python/qr/handler.py

Example:

curl -sS 'http://127.0.0.1:8080/qr?text=PythonQR'

Tip: write it to a file:

curl -sS 'http://127.0.0.1:8080/qr?text=PythonQR' --output qr.svg

`pack-qr` (Python QR, minimal packaging demo)¶

Route: /pack-qr
Methods: GET
Query: text
Content-Type: image/svg+xml
Source: examples/functions/python/pack-qr/handler.py

curl -sS 'http://127.0.0.1:8080/pack-qr?text=Hello' --output pack-qr.svg

`gmail-send`¶

Route: /gmail-send
Methods: GET, POST
Goal: Gmail SMTP helper for demos/integrations
Default behavior: dry_run=true (safe local testing without real credentials)
Source: examples/functions/python/gmail-send/handler.py

GET example:

curl -sS 'http://127.0.0.1:8080/gmail-send?to=demo@example.com&subject=Hi&text=Hello&dry_run=true'

Example response:

{"channel":"gmail","to":"demo@example.com","subject":"Hi","dry_run":true}

POST example:

curl -sS -X POST 'http://127.0.0.1:8080/gmail-send' \
  -H 'Content-Type: application/json' \
  -d '{"to":"demo@example.com","subject":"Hi","text":"Hello","dry_run":true}'

`sendgrid-send`¶

Route: /sendgrid-send
Methods: GET, POST
Goal: SendGrid email helper (safe by default)
Default behavior: dry_run=true
Env: SENDGRID_API_KEY (secret), SENDGRID_FROM
Source: examples/functions/python/sendgrid-send/handler.py

GET (dry run):

curl -sS 'http://127.0.0.1:8080/sendgrid-send?to=demo@example.com&subject=Hi&text=Hello&dry_run=true'

Example response (dry run):

{"function":"sendgrid-send","dry_run":true,"ok":true}

`sheets-webapp-append`¶

Route: /sheets-webapp-append
Methods: GET
Goal: append a row via a Google Apps Script Web App (safe by default)
Default behavior: dry_run=true
Env: SHEETS_WEBAPP_URL (secret)
Source: examples/functions/python/sheets-webapp-append/handler.py

Example (dry run):

curl -sS 'http://127.0.0.1:8080/sheets-webapp-append?sheet=Sheet1&values=a,b,c&dry_run=true'

`nombre`¶

Route: /nombre
Methods: GET (default policy)
Source: examples/functions/python/nombre/handler.py

Example:

curl -sS 'http://127.0.0.1:8080/nombre?name=Ana'

Example response:

{"runtime":"python","function":"hello","hello":"Ana","request_id":"<request_id>"}

`stripe-webhook-verify`¶

Route: /stripe-webhook-verify
Methods: POST
Goal: verify Stripe webhook signature (safe by default)
Default behavior: dry_run=true
Env: STRIPE_WEBHOOK_SECRET (secret)
Source: examples/functions/python/stripe-webhook-verify/handler.py

Dry run:

curl -sS -X POST 'http://127.0.0.1:8080/stripe-webhook-verify' \
  -H 'Content-Type: application/json' \
  -d '{"id":"evt_test"}'

Enforce verification:

set dry_run=false
include Stripe-Signature header
set STRIPE_WEBHOOK_SECRET in fn.env.json

`github-webhook-verify`¶

Route: /github-webhook-verify
Methods: POST
Goal: verify GitHub webhook signature (safe by default)
Default behavior: dry_run=true
Env: GITHUB_WEBHOOK_SECRET (secret)
Source: examples/functions/python/github-webhook-verify/handler.py

Node runtime¶

`hello@v2`¶

Route: /hello@v2
Methods: GET
Query: name
Source: examples/functions/node/hello/v2/handler.js
What it demonstrates: versioned endpoint (@v2) living in a version folder

Example:

curl -sS 'http://127.0.0.1:8080/hello@v2?name=NodeWay'

Typical response:

{"hello":"v2-NodeWay"}

`node-echo`¶

Route: /node-echo
Methods: GET, POST (editable in fn.config.json)
Query: name
Source: examples/functions/node/node-echo/handler.js

GET example:

curl -sS 'http://127.0.0.1:8080/node-echo?name=Node'

POST example:

curl -sS -X POST 'http://127.0.0.1:8080/node-echo?name=NodePost' \
  -H 'Content-Type: application/json' \
  -d '{}'

Example response:

{"runtime":"node","function":"node-echo","hello":"Node"}

`echo`¶

Route: /echo
Methods: GET
Query: key
Source: examples/functions/node/echo/handler.js
What it demonstrates: what FastFN passes in event.query and event.context.user

Example:

curl -sS 'http://127.0.0.1:8080/echo?key=test'

Typical response:

{"key":"test","query":{"key":"test"},"context":{"user":null}}

`qr@v2`¶

Route: /qr@v2
Methods: GET
Query: text or url, optional size
Content-Type: image/png
Source: examples/functions/node/qr/v2/handler.js

Example:

curl -sS 'http://127.0.0.1:8080/qr@v2?text=NodeQR' --output qr-node.png

`pack-qr-node` (Node QR, PNG output)¶

Route: /pack-qr-node
Methods: GET
Query: text
Content-Type: image/png
Source: examples/functions/node/pack-qr-node/handler.js

curl -sS 'http://127.0.0.1:8080/pack-qr-node?text=Hello' --output pack-qr-node.png

`ts-hello` (TypeScript handler)¶

Route: /ts-hello
Methods: GET
Query: name
Source: examples/functions/node/ts-hello/handler.ts

curl -sS 'http://127.0.0.1:8080/ts-hello?name=TypeScript'

`telegram-send`¶

Route: /telegram-send
Methods: GET, POST
Goal: Telegram Bot API helper for demos/integrations
Default behavior: dry_run=true (safe local testing without real credentials)
Source: examples/functions/node/telegram-send/handler.js
Layout: thin handler.js entrypoint delegating to a private core.js

GET example:

curl -sS 'http://127.0.0.1:8080/telegram-send?chat_id=123456&text=Hello&dry_run=true'

Example response (dry run):

{"channel":"telegram","chat_id":"123456","text":"Hello","dry_run":true}

POST example:

curl -sS -X POST 'http://127.0.0.1:8080/telegram-send' \
  -H 'Content-Type: application/json' \
  -d '{"chat_id":"123456","text":"Hello","dry_run":true}'

`telegram-ai-digest` (scheduled digest)¶

Route: /telegram-ai-digest
Methods: GET
Goal: fetch recent Telegram group messages, summarize with OpenAI, send digest back
Env (secrets): TELEGRAM_BOT_TOKEN, TELEGRAM_CHAT_ID, OPENAI_API_KEY
Source: examples/functions/node/telegram-ai-digest/handler.js
Layout: thin handler.js entrypoint delegating to a private core.js

Example:

curl -sS 'http://127.0.0.1:8080/telegram-ai-digest'

`telegram-ai-reply`¶

Route: /telegram-ai-reply
Methods: POST
Goal: Telegram webhook -> OpenAI -> Telegram reply
Env (secrets): TELEGRAM_BOT_TOKEN, OPENAI_API_KEY
Source: examples/functions/node/telegram-ai-reply/handler.js
Layout: thin handler.js entrypoint delegating to a private core.js
Deep dive: How telegram-ai-reply Works

Example (webhook POST):

curl -sS 'http://127.0.0.1:8080/telegram-ai-reply' \
  -X POST \
  -H 'Content-Type: application/json' \
  -d '{"message":{"chat":{"id":123},"text":"Hola"}}'

To send for real, set TELEGRAM_BOT_TOKEN and OPENAI_API_KEY in fn.env.json.

`whatsapp`¶

Route: /whatsapp
Methods: GET, POST, DELETE
Goal: WhatsApp real session manager (QR, connect, send, inbox/outbox, AI chat)
Source: examples/functions/node/whatsapp/handler.js
Layout: thin handler.js entrypoint delegating to a private core.js
Actions:
GET /whatsapp?action=qr
GET /whatsapp?action=status
POST /whatsapp?action=send
POST /whatsapp?action=chat
GET /whatsapp?action=inbox
DELETE /whatsapp?action=reset-session

Tip: if you call POST /whatsapp without an action, you may get 405 because some actions are intentionally method-specific.

`toolbox-bot` (safe tools runner)¶

Route: /toolbox-bot
Methods: GET, POST
Goal: parse tool directives ([[http:...]], [[fn:...]]) from text and execute them with allowlists
Source: examples/functions/node/toolbox-bot/handler.js
Layout: thin handler.js entrypoint delegating to a private core.js
Docs: Tools

Example:

curl -g -sS \
"http://127.0.0.1:8080/toolbox-bot?text=Use%20[[http:https://api.ipify.org?format=json]]%20and%20[[fn:hello|GET]]"

`ai-tool-agent` (OpenAI tool-calling agent)¶

Route: /ai-tool-agent
Methods: GET, POST
Goal: OpenAI chooses tools (http_get, fn_get) and the function executes them with allowlists
Default behavior: dry_run=true
Source: examples/functions/node/ai-tool-agent/handler.js
Layout: thin handler.js entrypoint delegating to a private core.js
Docs: Tools

Dry run:

curl -sS "http://127.0.0.1:8080/ai-tool-agent?dry_run=true&text=what%20is%20my%20ip%20and%20weather%20in%20Buenos%20Aires%3F"

Real run (requires OPENAI_API_KEY):

curl -sS "http://127.0.0.1:8080/ai-tool-agent?dry_run=false&text=what%20is%20my%20ip%20and%20weather%20in%20Buenos%20Aires%3F"

The response includes a trace.steps[] array with tool calls, tool results, and memory info.

Scheduler / cron:

ai-tool-agent ships with an example schedule block in examples/functions/node/ai-tool-agent/fn.config.json (disabled by default).
Enable schedules via the Console API, or by editing fn.config.json and reloading.
See: Manage Functions

`request-inspector`¶

Route: /request-inspector
Methods: GET, POST, PUT, PATCH, DELETE
Goal: show what the gateway passed into the handler (method/query/headers/body/context)
Source: examples/functions/node/request-inspector/handler.js
Layout: thin handler.js entrypoint delegating to a private core.js

Example:

curl -sS 'http://127.0.0.1:8080/request-inspector?key=test' \
  -X POST \
  -H 'x-demo: 1' \
  --data 'hello'

Edge / gateway patterns¶

These functions demonstrate a "Workers-like" pattern: validate the inbound request, rewrite it, then return a proxy directive.

They also follow the single-entry layout: each public handler.js stays tiny while a private sibling core.js holds the implementation logic.

`edge-proxy`¶

Route: /edge-proxy
Goal: minimal passthrough demo (proxies to /request-inspector)
Source: examples/functions/node/edge-proxy/handler.js

curl -sS 'http://127.0.0.1:8080/edge-proxy' | jq .

`edge-filter`¶

Route: /edge-filter
Goal: API key filter + rewrite + passthrough
Source: examples/functions/node/edge-filter/handler.js

curl -sS -i 'http://127.0.0.1:8080/edge-filter?user_id=123' | sed -n '1,12p'
curl -sS 'http://127.0.0.1:8080/edge-filter?user_id=123' -H 'x-api-key: dev' | jq .

`edge-auth-gateway`¶

Route: /edge-auth-gateway
Goal: Bearer auth gateway + passthrough
Source: examples/functions/node/edge-auth-gateway/handler.js

curl -sS -i 'http://127.0.0.1:8080/edge-auth-gateway?target=health' | sed -n '1,12p'
curl -sS 'http://127.0.0.1:8080/edge-auth-gateway?target=health' -H 'Authorization: Bearer dev-token' | jq .

`edge-header-inject`¶

Route: /edge-header-inject
Goal: inject headers and proxy to /request-inspector (so you can see them)
Source: examples/functions/node/edge-header-inject/handler.js

curl -sS 'http://127.0.0.1:8080/edge-header-inject?tenant=acme' -X POST --data 'hello' | jq .

`github-webhook-guard`¶

Route: /github-webhook-guard
Methods: POST
Goal: verify x-hub-signature-256 (GitHub HMAC) and optionally forward
Env: GITHUB_WEBHOOK_SECRET (secret)
Source: examples/functions/node/github-webhook-guard/handler.js

curl -sS -i 'http://127.0.0.1:8080/github-webhook-guard' \
  -X POST \
  -H 'x-hub-signature-256: sha256=bad' \
  --data '{"zen":"Keep it logically awesome.","hook_id":123}' | sed -n '1,12p'

`slack-webhook`¶

Route: /slack-webhook
Methods: GET
Goal: send a Slack Incoming Webhook (safe by default)
Default behavior: dry_run=true
Env: SLACK_WEBHOOK_URL (secret)
Source: examples/functions/node/slack-webhook/handler.js

Example (dry run):

curl -sS 'http://127.0.0.1:8080/slack-webhook?text=Hello&dry_run=true'

`discord-webhook`¶

Route: /discord-webhook
Methods: GET
Goal: send a Discord webhook (safe by default)
Default behavior: dry_run=true
Env: DISCORD_WEBHOOK_URL (secret)
Source: examples/functions/node/discord-webhook/handler.js

Example (dry run):

curl -sS 'http://127.0.0.1:8080/discord-webhook?content=Hello&dry_run=true'

`notion-create-page`¶

Route: /notion-create-page
Methods: GET
Goal: create a Notion page (safe by default)
Default behavior: dry_run=true
Env: NOTION_TOKEN (secret)
Source: examples/functions/node/notion-create-page/handler.js

Example (dry run):

curl -sS 'http://127.0.0.1:8080/notion-create-page?title=Hello&content=World&dry_run=true'

To send for real:

set dry_run=false
provide parent_page_id
set NOTION_TOKEN in fn.env.json

Autonomous Auto-install Examples¶

Python (inference)¶

auto-infer-no-requirements Source: examples/functions/python/auto-infer-no-requirements/handler.py
auto-infer-alias Source: examples/functions/python/auto-infer-alias/handler.py (#@requirements Pillow for PIL)
auto-infer-python-multi-deps Source: examples/functions/python/auto-infer-python-multi-deps/handler.py

Run:

curl -sS 'http://127.0.0.1:8080/auto-infer-no-requirements'
curl -sS 'http://127.0.0.1:8080/auto-infer-alias'
curl -sS 'http://127.0.0.1:8080/auto-infer-python-multi-deps'

Node (inference)¶

auto-infer-create-package Source: examples/functions/node/auto-infer-create-package/handler.js
auto-infer-update-package Source: examples/functions/node/auto-infer-update-package/handler.js
auto-infer-node-multi-deps Source: examples/functions/node/auto-infer-node-multi-deps/handler.js

Run:

curl -sS 'http://127.0.0.1:8080/auto-infer-create-package'
curl -sS 'http://127.0.0.1:8080/auto-infer-update-package'
curl -sS 'http://127.0.0.1:8080/auto-infer-node-multi-deps'

Tip:

- these multi-dependency examples are useful when you want to compare `native` inference with optional backends such as `pipreqs`, `detective`, or `require-analyzer`
- explicit manifests are still the recommended steady-state workflow

PHP (composer manifest)¶

auto-composer-basic Source: examples/functions/php/auto-composer-basic/handler.php
auto-composer-existing Source: examples/functions/php/auto-composer-existing/handler.php

Run:

curl -sS 'http://127.0.0.1:8080/auto-composer-basic'
curl -sS 'http://127.0.0.1:8080/auto-composer-existing'

Inspect resolution state:

curl -sS 'http://127.0.0.1:8080/_fn/function?runtime=node&name=auto-infer-create-package' \
| python3 - <<'PY'
import json,sys
obj=json.load(sys.stdin)
print(json.dumps((obj.get("metadata") or {}).get("dependency_resolution"), indent=2))
PY

Platform-Equivalent Advanced Patterns¶

Source package: examples/functions/platform-equivalents/

These examples mirror common patterns from Cloudflare/Vercel/Netlify/AWS guides, but in FastFN file-routing format.

Auth + RBAC profile¶

POST /auth/login (Node): issues HMAC bearer token
GET /auth/profile (Python): validates signature and expiration

TOKEN="$(curl -sS -X POST 'http://127.0.0.1:8080/auth/login' \
  -H 'content-type: application/json' \
  --data '{"username":"demo-admin","role":"admin"}' | \
  python3 -c 'import json,sys; print(json.load(sys.stdin)["token"])')"

curl -sS 'http://127.0.0.1:8080/auth/profile' \
  -H "authorization: Bearer ${TOKEN}"

Signed webhook + idempotency¶

POST /webhooks/github-signed (Python): verifies x-hub-signature-256, deduplicates x-github-delivery

PAYLOAD='{"action":"opened","repository":"fastfn"}'
SIG="$(python3 - <<'PY'
import hashlib,hmac
body=b'{"action":"opened","repository":"fastfn"}'
print("sha256=" + hmac.new(b"fastfn-webhook-secret", body, hashlib.sha256).hexdigest())
PY
)"
curl -sS -X POST 'http://127.0.0.1:8080/webhooks/github-signed' \
  -H "x-hub-signature-256: ${SIG}" \
  -H 'x-github-delivery: demo-1' \
  -H 'content-type: application/json' \
  --data "${PAYLOAD}"

Versioned orders API (polyglot)¶

POST /api/v1/orders (Python)
GET /api/v1/orders (Node)
GET /api/v1/orders/:id (PHP)
PUT /api/v1/orders/:id (Node)

ORDER_ID="$(curl -sS -X POST 'http://127.0.0.1:8080/api/v1/orders' \
  -H 'content-type: application/json' \
  --data '{"customer":"acme","items":[{"sku":"A-1","qty":2}]}' | \
  python3 -c 'import json,sys; print(json.load(sys.stdin)["order"]["id"])')"

curl -sS "http://127.0.0.1:8080/api/v1/orders/${ORDER_ID}"
curl -sS -X PUT "http://127.0.0.1:8080/api/v1/orders/${ORDER_ID}" \
  -H 'content-type: application/json' \
  --data '{"status":"shipped","tracking_number":"TRK-1001"}'

Background-like async job with polling¶

POST /jobs/render-report (Node): accepts work, returns 202
GET /jobs/render-report/:id (PHP): returns queued/running/succeeded

JOB_JSON="$(curl -sS -X POST 'http://127.0.0.1:8080/jobs/render-report' \
  -H 'content-type: application/json' \
  --data '{"report_type":"sales","items":[1,2,3,4]}')"
POLL_URL="$(printf '%s' "${JOB_JSON}" | python3 -c 'import json,sys; print(json.load(sys.stdin)["poll_url"])')"
curl -sS "http://127.0.0.1:8080${POLL_URL}"
sleep 3
curl -sS "http://127.0.0.1:8080${POLL_URL}"

PHP runtime¶

`php-profile`¶

Route: /php-profile
Methods: GET
Query: name
Source: examples/functions/php/php-profile/handler.php

Example:

curl -sS 'http://127.0.0.1:8080/php-profile?name=PHP'

Typical response:

{"runtime":"php","function":"php-profile","hello":"php-PHP"}

Rust runtime¶

`rust-profile`¶

Route: /rust-profile
Methods: GET
Query: name
Source: examples/functions/rust/rust-profile/lib.rs

Example:

curl -sS 'http://127.0.0.1:8080/rust-profile?name=Rust'

Typical response:

{"runtime":"rust","function":"rust-profile","hello":"rust-Rust"}

Control-plane endpoints (advanced)¶

Public example routes are things like /hello and /edge-filter. The platform control-plane lives under /_fn/* (OpenAPI, config, logs, reload).

In production, you normally restrict access to /_fn/* (or disable the console UI), and you should never let untrusted traffic reach it.

Reload discovery¶

After editing function files, refresh discovery:

curl -sS -X POST 'http://127.0.0.1:8080/_fn/reload'

Tail logs¶

Tail OpenResty or runtime logs (requires console API access):

curl -sS 'http://127.0.0.1:8080/_fn/logs?file=error&lines=200'
curl -sS 'http://127.0.0.1:8080/_fn/logs?file=access&lines=50&format=json'
curl -sS 'http://127.0.0.1:8080/_fn/logs?file=runtime&format=json&runtime=python&fn=hello&version=default&stream=stdout&lines=50' \
  -H 'x-fn-admin-token: my-secret-token'

Use file=runtime when you need the full handler debug stream. Response headers like X-Fn-Stdout are only a short preview.

Contract¶

Defines expected request/response shape, configuration fields, and behavioral guarantees.

End-to-End Example¶

Use the examples in this page as canonical templates for implementation and testing.

Edge Cases¶

Missing configuration fallbacks
Route conflicts and precedence
Runtime-specific nuances

Example Function Catalog¶

Run the examples¶

Beginner tour (10 minutes)¶

How these examples work¶

custom-handler-demo (custom handler name, Python + Node variants)¶

Multi-route demo apps¶

next-style (recommended: Next.js-style file routing)¶

polyglot-tutorial (step-by-step multi-runtime pipeline)¶

polyglot-db-demo (shared SQLite across runtimes)¶

ip-intel (file routes + optional deps + deterministic mock mode)¶

Python runtime¶

cron-tick (scheduler demo)¶

utc-time (cron + timezone demo)¶

offset-time (cron + timezone demo)¶

tools-loop (tools loop demo, inspired by agent loops)¶

telegram-ai-reply-py (Telegram AI bot, Python)¶

hello¶

risk-score¶

slow¶

html-demo¶

csv-demo¶

png-demo¶

lambda-echo¶

custom-echo¶

requirements-demo¶

qr¶

pack-qr (Python QR, minimal packaging demo)¶

gmail-send¶

sendgrid-send¶

sheets-webapp-append¶

nombre¶

stripe-webhook-verify¶

github-webhook-verify¶

Node runtime¶

hello@v2¶

node-echo¶

echo¶

qr@v2¶

pack-qr-node (Node QR, PNG output)¶

ts-hello (TypeScript handler)¶

telegram-send¶

telegram-ai-digest (scheduled digest)¶

telegram-ai-reply¶

whatsapp¶

toolbox-bot (safe tools runner)¶

ai-tool-agent (OpenAI tool-calling agent)¶

request-inspector¶

Edge / gateway patterns¶

edge-proxy¶

edge-filter¶

edge-auth-gateway¶

edge-header-inject¶

github-webhook-guard¶

slack-webhook¶

discord-webhook¶

notion-create-page¶

Autonomous Auto-install Examples¶

Python (inference)¶

Node (inference)¶

PHP (composer manifest)¶

Platform-Equivalent Advanced Patterns¶

Auth + RBAC profile¶

Signed webhook + idempotency¶

Versioned orders API (polyglot)¶

Background-like async job with polling¶

PHP runtime¶

php-profile¶

Rust runtime¶

rust-profile¶

Control-plane endpoints (advanced)¶

Reload discovery¶

Tail logs¶

Contract¶

End-to-End Example¶

Edge Cases¶

See also¶

`custom-handler-demo` (custom handler name, Python + Node variants)¶

`next-style` (recommended: Next.js-style file routing)¶

`polyglot-tutorial` (step-by-step multi-runtime pipeline)¶

`polyglot-db-demo` (shared SQLite across runtimes)¶

`ip-intel` (file routes + optional deps + deterministic mock mode)¶

`cron-tick` (scheduler demo)¶

`utc-time` (cron + timezone demo)¶

`offset-time` (cron + timezone demo)¶

`tools-loop` (tools loop demo, inspired by agent loops)¶

`telegram-ai-reply-py` (Telegram AI bot, Python)¶

`hello`¶

`risk-score`¶

`slow`¶

`html-demo`¶

`csv-demo`¶

`png-demo`¶

`lambda-echo`¶

`custom-echo`¶

`requirements-demo`¶

`qr`¶

`pack-qr` (Python QR, minimal packaging demo)¶

`gmail-send`¶

`sendgrid-send`¶

`sheets-webapp-append`¶

`nombre`¶

`stripe-webhook-verify`¶

`github-webhook-verify`¶

`hello@v2`¶

`node-echo`¶

`echo`¶

`qr@v2`¶

`pack-qr-node` (Node QR, PNG output)¶

`ts-hello` (TypeScript handler)¶

`telegram-send`¶

`telegram-ai-digest` (scheduled digest)¶

`telegram-ai-reply`¶

`whatsapp`¶

`toolbox-bot` (safe tools runner)¶

`ai-tool-agent` (OpenAI tool-calling agent)¶

`request-inspector`¶

`edge-proxy`¶

`edge-filter`¶

`edge-auth-gateway`¶

`edge-header-inject`¶

`github-webhook-guard`¶

`slack-webhook`¶

`discord-webhook`¶

`notion-create-page`¶

`php-profile`¶

`rust-profile`¶