Catch, log, and inspect every HTTP request — self-hosted, zero dependencies, and yours forever.
Testing webhook integrations is painful. You guess what your service sent, hit F5 a hundred times, and pray the request format is right. Kapture gives you a dead-simple endpoint that logs everything — method, headers, body, query params, IP — and lets you inspect it in a clean UI.
But here's the thing about tools like webhook.site and RequestBin: every webhook you send to them leaves your machine. Your payloads live on someone else's server, they expire in hours, and if you need to check what Stripe sent you last week — it's gone.
Kapture is self-hosted, open-source, and persists your logs as long as you need them. One command and your data stays yours forever. No rate limits, no signup, no "try again tomorrow." Zero dependencies — drop it on any server (even shared hosting) and it just works.
And now it also forwards. Set FORWARD_URL and Kapture becomes a transparent proxy: capture every request your backend receives, inspect it in the dashboard, and forward it unchanged — all in a single pass. Debug webhooks without taking your integration offline.
php -S 0.0.0.0:8080 -t publicSend a webhook:
curl -X POST -H "Content-Type: application/json" \
-d '{"event":"user.created","email":"user@example.com"}' \
http://localhost:8080/kapture/anything-you-likeOpen the admin:
open http://localhost:8080/adminPassword: changeme
To switch to SQLite storage, set STORAGE_DRIVER=sqlite in .env (requires ext-sqlite3).
- Self-hosted — your webhooks never leave your machine. No third party sees your payloads.
- Persistent logs — stays across restarts. Check what Shopify sent you last week? Open the archive.
- Configurable retention — keep logs for 7 days or 7 months. Prune when you're ready.
- Zero dependencies — just PHP 8.4+. No Composer install, no database setup, no Docker required.
- One command to run —
php -S 0.0.0.0:8080 -t publicand you're capturing. - Works everywhere — laptop, shared hosting, DigitalOcean box, Raspberry Pi.
- Full capture — headers, body, query params, IP, timestamp, any HTTP method
- Dark theme admin UI — expandable details, text filter, live auto-refresh
- Clickable URL parts — click any path segment or
?param=valueto instantly filter the table by that value. Each query param key gets its own color so you can spot patterns at a glance. - Archive browser — pick any daily log from the sidebar. Browse tomorrow what came in today.
- Raw dump —
?rawfor JSONL access. Pipe intojq, grep, or your own tooling.
Set FORWARD_URL and Kapture stops being just a sink — it becomes a transparent proxy sitting between your webhook provider and your real backend.
How it works: Stripe sends a webhook → Kapture captures it, then instantly forwards it to your actual endpoint. You see the full request in the dashboard, your backend receives it unchanged. No integration code. No duplicate endpoints. One config variable.
Stripe/GitHub/Shopify ──▶ Kapture ──▶ Your real server
capture + log unchanged
Why this matters:
- Zero downtime debugging — your backend keeps running while you inspect every request that hits it
- See what actually arrived — not what you think arrived. Catch malformed payloads, missing headers, wrong content types before they break your code
- Color-coded status badges — forwarded requests show their response status on the dashboard. Green for 200, orange for 429, red for 500. Spot failures at a glance
- No tunnel required — runs on your existing server. No ngrok, no cloudflare, no third-party relay
To enable, add one line to your .env:
FORWARD_URL=https://your-real-server.com/webhook
Kapture captures the request, forwards it to your server, logs the response status, and stores the result — all in a single pass. Your server never knows Kapture was there.
- JSONL files — one JSON object per line, standard format, readable by any tool
- SQLite option — set
STORAGE_DRIVER=sqlitefor a single database file - Daily rotation — automatic, with configurable pruning
- Self-hosted — no third-party server sees your payloads
- Persistent logs — stays across restarts, browsable by day, configurable retention
- Forwarding proxy — capture AND forward in one pass. Set
FORWARD_URL, inspect every request your backend receives - Zero dependencies — just PHP. No Composer, no Docker, no database setup
- No rate limits, no signup — run it, use it, done
- Runs anywhere — laptop, shared hosting, VPS, Raspberry Pi
POST /kapture/your-custom-path
GET /kapture/anything?foo=bar
PUT /kapture/test
...
Returns {"ok":true,"captureId":"<unique-id>"}. Keep the capture ID to find it in the logs.
Set FORWARD_URL in .env and every captured request is also forwarded to your real backend. The forwarded request's response status code is logged alongside the captured data, color-coded in the dashboard:
- Green — 2xx success
- Orange — 4xx client error
- Red — 5xx server error
Your backend receives the original request unchanged. Kapture sits silently in between, capturing and forwarding in one pass.
| URL | What |
|---|---|
/admin |
UI — browse, filter, expand requests |
/admin?raw |
Raw JSONL dump of today's file |
/admin?file=2026-05-23 |
Browse a specific day's log |
/admin?file=2026-05-23&raw |
Raw dump of a specific day |
Pro tip: in the table, click any URI segment (/stripe, /webhook) or query parameter (?event=created, ?source=shopify) to filter all matching entries. Click again to clear. Each query param key has a distinct pastel color — spot patterns at a glance.
Copy .env.example → .env and edit — the app won't start without it:
ADMIN_PASSWORD=changeme # Admin login password
LOG_DIR=./logs # Where logs / database are stored
ROTATE_DAYS=7 # Days to keep logs (filesystem only)
STORAGE_DRIVER=filesystem # 'filesystem' (default) or 'sqlite'
FORWARD_URL= # Optional: forward captured requests to this URLEach request is logged as a single JSON line (JSONL):
{
"capturedAt": "2026-05-23T14:30:00Z",
"method": "POST",
"uri": "/kapture/orders",
"query": {"source": "shopify"},
"headers": {
"Content-Type": "application/json",
"User-Agent": "Shopify-Captain-Hook/1.0"
},
"body": "{\"event\":\"order.created\"}",
"ip": "203.0.113.42",
"captureId": "a1b2c3d4e5f6g7h8",
"forwardUrl": "https://your-server.com/webhook",
"forwardStatusCode": 200
}- PHP 8.4+
php -S 0.0.0.0:8080 -t publicThe built-in server handles routing. No Apache or Nginx needed.
composer testcomposer check # phpstan + ecs + tests
composer ecs:fix # auto-fix code style- PHPStan — static analysis at level 8
- ECS — PSR-12 code style with Symfony rules
- PHPUnit
Does it work with Stripe / Shopify / GitHub webhooks? Yes. Kapture accepts any HTTP method and captures the full request — headers, body, query params, and IP. Drop the endpoint URL into any webhook provider's dashboard.
Can I run it in production? Kapture is designed for local development and testing. For production use, add HTTPS and a stronger password.
Does it persist logs between restarts?
Yes. Logs are written to logs/ as JSONL files (or a single kapture.db SQLite file when using STORAGE_DRIVER=sqlite). The admin UI lets you browse any daily archive from the sidebar.
How is this different from webhook.site? webhook.site is convenient for one-off testing, but your webhooks go through their servers, expire quickly, and you can't browse yesterday's data. Kapture is self-hosted, open-source, and keeps your logs as long as you configure it to. Your data never leaves your machine. It's not a temporary buffer — it's your webhook archive.
What if I need to check a webhook from last week?
Open /admin?file=2026-05-20 and scroll. Kapture persists logs across restarts, organized by day, browsable from the sidebar. Set ROTATE_DAYS=90 and you have a 3-month audit trail.
What is the forwarding proxy mode?
Set FORWARD_URL in your .env and Kapture becomes a transparent proxy: it captures every incoming request, then immediately forwards it to your real backend. Your server receives the request unchanged — Kapture just silently logs the response status code and stores it alongside the captured data. Think of it as a webhook inspection layer that sits between Stripe and your app, letting you see everything without interrupting your integration.
- Webhook forwarding / proxy mode
- Docker image for one-command deploy
- Webhook replay — resend captured requests on demand
- Configurable log retention per route
- CLI tail command for live log streaming
Contributions are welcome! Here's how to help:
- Report a bug
- Submit a pull request
- Star the repo to show support ⭐
MIT