Wick 0.7: Local HTTP API

March 26, 2026

Until now, Wick only worked through MCP — which meant only Claude Code and Cursor could use it. That was fine for the initial use case, but it left out everyone writing Python scripts, building LangChain agents, wiring up n8n workflows, or doing anything outside the MCP ecosystem.

Wick 0.7 fixes that. One command:

wick serve --api

That starts a local HTTP API server on localhost:8090. Now anything on your machine can make browser-grade web requests.

Endpoints

GET /v1/fetch?url=https://nytimes.com          → page as markdown
GET /v1/crawl?url=https://docs.example.com      → multi-page crawl
GET /v1/map?url=https://example.com              → discover all URLs
GET /v1/search?q=rust+async                      → web search
GET /health                                      → status check

All endpoints return JSON. Same Chrome TLS fingerprint, same 100% anti-bot success rate, same clean markdown output — just accessible via HTTP instead of MCP.

Examples

Python:

import requests

r = requests.get("http://localhost:8090/v1/fetch",
    params={"url": "https://nytimes.com"})
data = r.json()
print(data["content"])    # markdown
print(data["title"])      # page title
print(data["timing_ms"])  # fetch time

JavaScript:

const r = await fetch(
  "http://localhost:8090/v1/crawl?url=https://docs.example.com&max_pages=5"
);
const { pages } = await r.json();
pages.forEach(p => console.log(p.title));

curl:

curl -s "http://localhost:8090/v1/map?url=https://getwick.dev" | jq '.urls[]'

Why local?

Every other web access API (Firecrawl, Bright Data, Scrapfly) runs in the cloud. Your requests go through their servers. You pay per request. Your data leaves your machine.

Wick's API runs on localhost. Your data never leaves your network. The requests exit from your residential IP. There's no per-request billing — Pro is $20/month flat, unlimited requests. And there's no cloud dependency — if their servers go down, you don't care.

What this enables

RAG pipelines — Crawl documentation sites and feed them into your vector store
LangChain / LlamaIndex — Use Wick as your web loader tool
n8n / Zapier — HTTP request node pointed at localhost
Jupyter notebooks — Research and data collection
Custom agents — Any framework, any language
Shell scripts — curl + jq and you're scraping

Install / Upgrade

# Fresh install
brew tap wickproject/wick && brew install wick

# Upgrade
brew upgrade wick

# Start the API
wick serve --api

Previous releases

GitHub | API Docs | Contact