HTTP API

Squasher’s ingestion API accepts the Squasher event schema over HTTP. No SDK required — if you can make an HTTP request that sends structured events, you can send data to Squasher. This is the universal fallback for any language, framework, or platform that doesn’t have a dedicated SDK. If you use the Log Connectors screen in Squasher, create the HTTP API connector first and copy the managed endpoint and ingest key from there. The endpoint below uses the same wire format.

Endpoint

POST https://ingest.squasher.ai/v1/ingest/{project_id}

Authentication

Header	Value
`x-squasher-key`	Your project API key (`sq_pk_...`)
`Content-Type`	`application/json` or `application/x-ndjson`

Quick Start

Send a test event using curl:

curl -X POST "https://ingest.squasher.ai/v1/ingest/YOUR_PROJECT_ID" \
  -H "Content-Type: application/json" \
  -H "x-squasher-key: sq_pk_your_api_key" \
  -d '{
    "message": "Hello from Squasher!",
    "level": "info",
    "environment": "production",
    "tags": { "service": "my-app" }
  }'

Response: 202 Accepted

{
  "id": "evt_550e8400-e29b-41d4-a716-446655440000",
  "status": "accepted"
}

Batch Formats

Send one JSON object for a single event. To send multiple events, use any of these formats:

A top-level JSON array
An object containing an events array
Newline-delimited JSON (NDJSON), with one object per non-empty line

For an events envelope:

{
  "events": [
    { "message": "Order created", "level": "info" },
    { "message": "Payment failed", "level": "error", "type": "PaymentError" }
  ]
}

The equivalent top-level array is:

[
  { "message": "Order created", "level": "info" },
  { "message": "Payment failed", "level": "error", "type": "PaymentError" }
]

For NDJSON, set Content-Type: application/x-ndjson:

{"message":"Order created","level":"info"}
{"message":"Payment failed","level":"error","type":"PaymentError"}

Empty bodies, empty batches, malformed JSON, and non-object events return 400 Bad Request and are not forwarded.

Payload Fields

Field	Type	Required	Description
`message`	string	Yes	Log message or error description
`type`	string	No	Error class name (e.g. `TypeError`, `ConnectionError`)
`level`	enum	No	`fatal`, `error`, `warning`, `info`, `debug`. Default: `error`
`stack`	string	No	Raw stack trace string
`frames`	array	No	Parsed stack frames (preferred over raw `stack`)
`environment`	string	No	Environment tag (e.g. `production`, `staging`)
`release`	string	No	Release/version string (e.g. `v1.2.3`)
`tags`	object	No	Arbitrary key-value metadata
`user`	object	No	User context: `id`, `email`, `username`, `ip_address`
`request`	object	No	Request context: `url`, `method`, `headers`
`breadcrumbs`	array	No	Trail of events leading to the error
`extra`	object	No	Arbitrary extra data
`timestamp`	string	No	ISO 8601 timestamp. Default: server receive time

Log Levels

Squasher processes all log levels but treats them differently:

Level	Indexed for Search	Creates Error Group	Triggers Alerts	Counts Against
`fatal`	Yes	Yes	Yes	Error quota
`error`	Yes	Yes	Yes	Error quota
`warning`	Yes	No	No	Log quota
`info`	Yes	No	No	Log quota
`debug`	Yes	No	No	Log quota

Non-error events provide context for AI triage — when an error occurs, Squasher uses surrounding log entries to build a richer diagnosis.

Examples

Error with Stack Trace

curl -X POST "https://ingest.squasher.ai/v1/ingest/YOUR_PROJECT_ID" \
  -H "Content-Type: application/json" \
  -H "x-squasher-key: sq_pk_your_api_key" \
  -d '{
    "message": "Cannot read properties of undefined",
    "type": "TypeError",
    "level": "error",
    "stack": "TypeError: Cannot read properties of undefined\n    at UserList (src/UserList.tsx:24:18)",
    "frames": [
      {
        "filename": "src/UserList.tsx",
        "function": "UserList",
        "lineno": 24,
        "colno": 18,
        "in_app": true
      }
    ],
    "environment": "production",
    "release": "v1.2.3",
    "user": { "id": "user_42", "email": "[email protected]" }
  }'

Structured Log Entry

curl -X POST "https://ingest.squasher.ai/v1/ingest/YOUR_PROJECT_ID" \
  -H "Content-Type: application/json" \
  -H "x-squasher-key: sq_pk_your_api_key" \
  -d '{
    "message": "Order processed successfully",
    "level": "info",
    "tags": {
      "order_id": "ord_12345",
      "amount": 49.99,
      "currency": "USD"
    }
  }'

Python

import requests

requests.post(
    "https://ingest.squasher.ai/v1/ingest/YOUR_PROJECT_ID",
    headers={
        "Content-Type": "application/json",
        "x-squasher-key": "sq_pk_your_api_key",
    },
    json={
        "message": "Database connection failed",
        "type": "ConnectionError",
        "level": "error",
        "tags": {"db_host": "db.example.com", "retry_count": 3},
    },
)

Go

payload := map[string]interface{}{
    "message": "Request timeout exceeded",
    "type":    "TimeoutError",
    "level":   "error",
    "tags":    map[string]interface{}{"endpoint": "/api/users", "duration_ms": 30000},
}
body, _ := json.Marshal(payload)

req, _ := http.NewRequest("POST",
    "https://ingest.squasher.ai/v1/ingest/YOUR_PROJECT_ID", bytes.NewReader(body))
req.Header.Set("Content-Type", "application/json")
req.Header.Set("x-squasher-key", "sq_pk_your_api_key")

http.DefaultClient.Do(req)

Ruby

require "net/http"
require "json"

uri = URI("https://ingest.squasher.ai/v1/ingest/YOUR_PROJECT_ID")
req = Net::HTTP::Post.new(uri, {
  "Content-Type" => "application/json",
  "x-squasher-key" => "sq_pk_your_api_key",
})
req.body = { message: "Unexpected nil value", type: "NoMethodError", level: "error" }.to_json
Net::HTTP.start(uri.hostname, uri.port, use_ssl: true) { |http| http.request(req) }

Rate Limits

Edge: 1 MB max payload per request for SDK/drain routes
OTLP: 5 MB max payload for /v1/traces and /v1/logs
Project admission limit — batches that exceed max_events_per_minute return 429 with Retry-After and X-RateLimit-* headers; retry after the advertised reset

Combining with SDKs

The HTTP API accepts the same payload format as Squasher’s SDKs. You can mix and match — use the SDK for your primary language and the HTTP API for auxiliary services. All events appear in the same dashboard and are deduplicated by fingerprint.

AI batch ingest

For LLM generations, agent sessions, and tool calls, prefer the Agent telemetry SDK. If you need direct HTTP, send Langfuse-compatible AI batches to POST /v1/ai/ingest/{project_id}. The OpenRouter integration has a complete curl payload.

Agent handoff

Use Squasher HTTP ingest only when no first-party SDK or OTLP exporter fits. Read SQUASHER_API_KEY and SQUASHER_PROJECT_ID from environment variables, send one safe verification event, then confirm it appears with logs or errors search. Do not hardcode real API keys.

Get Started

SDKs

Integrations

Migrations

Cloud Log Drains

Features

API Reference

Endpoint

Authentication

Quick Start

Batch Formats

Payload Fields

Log Levels

Examples

Error with Stack Trace

Structured Log Entry

Python

Go

Ruby

Rate Limits

Combining with SDKs

AI batch ingest

Agent handoff

​Endpoint

​Authentication

​Quick Start

​Batch Formats

​Payload Fields

​Log Levels

​Examples

​Error with Stack Trace

​Structured Log Entry

​Python

​Go

​Ruby

​Rate Limits

​Combining with SDKs

​AI batch ingest

​Agent handoff

Endpoint

Authentication

Quick Start

Batch Formats

Payload Fields

Log Levels

Examples

Error with Stack Trace

Structured Log Entry

Python

Go

Ruby

Rate Limits

Combining with SDKs

AI batch ingest

Agent handoff