Documentation

Overview Quick start Node.js SDK Python SDK OpenAI Anthropic Gemini LangChain Claude MCP REST API x402 — Pay per log Log fields

Logwick Documentation

Audit logging for AI agents

Everything you need to log, search, and understand what your AI agents are doing in production. One line of code. Full visibility.

Get your API key →

Overview

Logwick is an audit log for AI agents. After each AI call in your code, you send one POST request to Logwick with the input, output, agent, and status. Logwick stores it, indexes it, and makes it searchable from your dashboard.

Use it to debug production issues, monitor costs and error rates, meet compliance requirements, and understand what your AI agents are actually doing.

REST API

POST logs from any language or framework

Node.js SDK

npm install logwick

Python SDK

pip install logwick

Claude MCP

Ask Claude about your logs in plain English

Dashboard

Search, filter, export, and set webhooks

Webhooks

Get alerted when error rates spike

Quick start

Get up and running in under 3 minutes. No SDK required — just a fetch call.

1. Get your API key

2. Add one call after your AI request

JavaScript — add after any AI call

const start = Date.now()
const result = await openai.chat.completions.create({
  model: 'gpt-4o',
  messages: [{ role: 'user', content: userPrompt }]
})

// Add this — fire and forget, never blocks your code
fetch('https://logwick.io/api/v1/logs', {
  method: 'POST',
  headers: {
    'Authorization': 'Bearer sk-lw-your-key',
    'Content-Type': 'application/json'
  },
  body: JSON.stringify({
    agent:      'gpt-4o',
    action:     'email_draft',
    status:     'success',
    input:      userPrompt,
    output:     result.choices[0].message.content,
    tokens:     result.usage.total_tokens,
    latency_ms: Date.now() - start,
    user:       req.user.email,
  })
}).catch(() => {}) // never throws

3. Check your dashboard

Open logwick.io/dashboard — your log appears instantly. That's it.

Node.js SDK

Install

npm install logwick

Basic usage

JavaScript

import { LogwickClient } from 'logwick'

const logwick = new LogwickClient({ apiKey: 'sk-lw-your-key' })

// Fire and forget
logwick.fire({
  agent:  'gpt-4o',
  action: 'email_draft',
  status: 'success',
  input:  userPrompt,
  output: result.choices[0].message.content,
  tokens: result.usage.total_tokens,
  user:   currentUser.email,
})

Client options

JavaScript

const logwick = new LogwickClient({
  apiKey:  'sk-lw-your-key',  // required
  silent:  true,               // suppress console warnings (default: true)
  tags:    ['production'],     // default tags on every log
})

Python SDK

Install

pip install logwick

Basic usage

Python

import logwick

logwick.init(api_key="sk-lw-your-key")

# Fire and forget
logwick.fire({
    "agent":  "gpt-4o",
    "action": "email_draft",
    "status": "success",
    "input":  user_prompt,
    "output": result,
    "tokens": 312,
    "user":   user_email,
})

Using the client directly

Python

from logwick import LogwickClient

lw = LogwickClient(
    api_key="sk-lw-your-key",
    silent=False,            # print warnings
    tags=["production"],     # default tags
)

OpenAI integration

Wrap your OpenAI call and Logwick automatically captures input, output, tokens, cost, and latency.

Node.js

JavaScript

const result = await logwick.openai(
  () => openai.chat.completions.create({
    model: 'gpt-4o',
    messages: [{ role: 'user', content: prompt }]
  }),
  { action: 'email_draft', user: req.user.email }
)
// result is the normal OpenAI response — nothing changes

Python

result = lw.openai(
    lambda: client.chat.completions.create(
        model="gpt-4o",
        messages=[{"role": "user", "content": prompt}]
    ),
    {"action": "email_draft", "user": user_email}
)

Anthropic / Claude integration

Wrap your Anthropic messages call and Logwick captures everything automatically.

Node.js

JavaScript

const result = await logwick.anthropic(
  () => anthropic.messages.create({
    model: 'claude-3-5-sonnet-20241022',
    max_tokens: 1024,
    messages: [{ role: 'user', content: prompt }]
  }),
  { action: 'document_review', user: req.user.email }
)

Python

result = lw.anthropic(
    lambda: client.messages.create(
        model="claude-3-5-sonnet-20241022",
        max_tokens=1024,
        messages=[{"role": "user", "content": prompt}]
    ),
    {"action": "document_review", "user": user_email}
)

Google Gemini integration

Node.js

JavaScript

const result = await logwick.gemini(
  () => model.generateContent(prompt),
  { action: 'data_analysis', user: req.user.email }
)

Python

result = lw.gemini(
    lambda: model.generate_content(prompt),
    {"action": "data_analysis", "user": user_email}
)

LangChain integration

Add one callback handler and every LLM call in your chain is logged automatically — no per-call code needed.

Node.js

JavaScript

import { LogwickCallbackHandler } from 'logwick'

const handler = new LogwickCallbackHandler(logwick, {
  user: 'ops@acme.com'
})

const chain = new LLMChain({
  llm,
  prompt,
  callbacks: [handler]  // every call in the chain is now logged
})

Python

handler = lw.langchain_handler(user="ops@acme.com")

chain = LLMChain(
    llm=llm,
    prompt=prompt,
    callbacks=[handler]  # every call in the chain is now logged
)

Claude MCP integration

Connect Logwick to Claude Desktop and ask questions about your logs in plain English. This is the fastest way to investigate AI agent incidents.

"Show me my last 10 error logs"
"What was my success rate this week?"
"How many tokens did I spend yesterday?"
"Find all failed email_draft actions"
"Log this GPT-4o call for me"

Setup with Claude Desktop

Add this to your claude_desktop_config.json and restart Claude Desktop.

Mac: ~/Library/Application Support/Claude/claude_desktop_config.json

Windows: %APPDATA%\Claude\claude_desktop_config.json

claude_desktop_config.json

{
  "mcpServers": {
    "logwick": {
      "command": "npx",
      "args": ["-y", "@logwick/mcp"],
      "env": {
        "LOGWICK_API_KEY": "sk-lw-your-key"
      }
    }
  }
}

Available tools

ingest_logWrite a log entry to Logwick

query_logsSearch logs by status, agent, action, date, or keyword

get_statsGet usage stats — success rate, tokens, cost

get_logGet full details of a single log entry by ID

delete_logDelete a log entry

Adding Logwick to a project via Claude

Once the MCP server is connected, you can ask Claude to add Logwick to any project:

Say this to Claude

"Go to https://logwick.io/docs and add Logwick logging 
to my project. My API key is sk-lw-your-key."

Claude will read this documentation, install the appropriate SDK, add your API key to your environment variables, and wire up logging to your existing AI calls.

REST API

All endpoints require an API key in the Authorization header.

Authentication

Authorization: Bearer sk-lw-your-key

POST /api/v1/logs — Ingest a log

Request

curl -X POST https://logwick.io/api/v1/logs \
  -H "Authorization: Bearer sk-lw-your-key" \
  -H "Content-Type: application/json" \
  -d '{
    "agent":      "gpt-4o",
    "action":     "email_draft",
    "status":     "success",
    "input":      "Draft a follow-up email",
    "output":     "Subject: Following up...",
    "tokens":     312,
    "latency_ms": 1842,
    "user":       "customer@acme.com"
  }'

Response

{
  "id": "fcf559c2-a3cb-4d48-999c-b606f1440472",
  "timestamp": "2026-04-19T13:44:52.241063+00:00",
  "status": "ingested"
}

GET /api/v1/logs — Query logs

Request

# All logs
curl https://logwick.io/api/v1/logs \
  -H "Authorization: Bearer sk-lw-your-key"

# Filter by status
curl "https://logwick.io/api/v1/logs?status=error" \
  -H "Authorization: Bearer sk-lw-your-key"

# Filter by agent
curl "https://logwick.io/api/v1/logs?agent=gpt-4o" \
  -H "Authorization: Bearer sk-lw-your-key"

# Search
curl "https://logwick.io/api/v1/logs?search=email" \
  -H "Authorization: Bearer sk-lw-your-key"

# Export as CSV
curl "https://logwick.io/api/v1/logs?format=csv" \
  -H "Authorization: Bearer sk-lw-your-key"

GET /api/v1/stats — Usage statistics

Request

curl "https://logwick.io/api/v1/stats?days=30" \
  -H "Authorization: Bearer sk-lw-your-key"

Response

{
  "total": 1842,
  "success": 1756,
  "error": 86,
  "success_rate": 95.4,
  "error_rate": 4.6,
  "avg_latency": 1204,
  "total_tokens": 284921,
  "total_cost": 4.27,
  "period_days": 30
}

x402 — Pay per log

Logwick supports x402 — an open payment protocol that lets AI agents pay per log with USDC on Base. No account, no API key, no signup required. The agent pays and logs in a single request.

Endpoint

POST https://logwick.io/api/v1/agent-log

Price: $0.001 USDC per log · Network: Base (eip155:8453) · No API key required

How it works

The agent calls the endpoint. Logwick responds with HTTP 402 and payment requirements. The agent pays $0.001 USDC, includes the payment proof in the request header, and Logwick stores the log.

Discover pricing — GET request

curl https://logwick.io/api/v1/agent-log

# Returns:
# {
#   "x402Version": 1,
#   "accepts": [{
#     "scheme": "exact",
#     "network": "eip155:8453",
#     "maxAmountRequired": "1000",
#     "payTo": "0x...",
#     "description": "Ingest one AI agent audit log entry"
#   }]
# }

Log a paid entry

POST with x402 payment header

curl -X POST https://logwick.io/api/v1/agent-log   -H "Content-Type: application/json"   -H "X-Payment: <signed-payment-proof>"   -d '{
    "agent":      "gpt-4o",
    "action":     "email_draft",
    "status":     "success",
    "input":      "Draft a follow-up email",
    "output":     "Subject: Following up...",
    "tokens":     312,
    "latency_ms": 1842
  }'

Using the Coinbase AgentKit

If your agent uses Coinbase AgentKit, x402 payments are handled automatically — the agent discovers the price, pays, and logs in one step.

JavaScript — AgentKit with x402

import { CdpWalletProvider } from '@coinbase/agentkit'
import { x402Fetch } from '@coinbase/x402-fetch'

const wallet = await CdpWalletProvider.configureWithWallet({ /* config */ })

// x402Fetch automatically handles payment — no manual signing needed
const response = await x402Fetch('https://logwick.io/api/v1/agent-log', {
  method: 'POST',
  headers: { 'Content-Type': 'application/json' },
  body: JSON.stringify({
    agent:      'gpt-4o',
    action:     'email_draft',
    status:     'success',
    input:      prompt,
    output:     result,
    tokens:     312,
  }),
  wallet,
})

Log fields

Same fields as the standard REST API — see Log fields reference below. The only difference is authentication — payment via x402 instead of an API key.

✓ No account required — payment is authentication

✓ $0.001 per log — fractions of a cent, agent-friendly pricing

✓ Base mainnet — real USDC, instant settlement

✓ Listed on agentic.market — discoverable by any x402-compatible agent

Log fields reference

All fields accepted by the ingest endpoint. Fields marked with * are required.

FieldTypeDescription

agent*

string

The AI model or agent name (e.g. gpt-4o, claude-3-5-sonnet, gemini-1.5-pro)

action*

string

The action or task type (e.g. email_draft, data_analysis, code_generation)

status

string

success, error, or pending (default: success)

input

string

The prompt or input sent to the AI agent

output

string

The response or output from the AI agent

user

string

The user or system that triggered this action

tokens

number

Total tokens used (input + output)

latency_ms

number

Response time in milliseconds

cost_usd

number

Estimated cost in USD