Firecrawl proxy setup
A self-hosted Firecrawl reads its proxy from three variables in `.env`, and its browser service picks them up too. Point them at our HTTP port with your login.
- Protocols
- HTTP proxy on port 7000 with a login
- Authentication
PROXY_USERNAMEandPROXY_PASSWORDin.env- Rotation
- A new IP per connection with a plain login; one IP for the whole instance with a
-sid-key - Works with
- Residential and mobile proxies, on a self-hosted Firecrawl
Before you start
- Firecrawl running from its repository with Docker Compose, as its self-hosting guide describes.
- Your proxy login and password from the Residential or Mobile page of the dashboard.
- Traffic on the product you use: a login without traffic gets 402.
- Host
gate.aeonproxy.com- Port (HTTP / HTTPS)
7000- Login
- your proxy login, with targeting keys if you need them:
LOGIN-cc-us-sid-abc123 - Password
- your proxy password
Copy the login and password from the Residential or Mobile page of the dashboard; the generator there also writes ready lines. Key syntax: targeting and sessions.
Set the proxy
In the .env file next to the Compose file, fill in the proxy section:
PROXY_SERVER=http://gate.aeonproxy.com:7000
PROXY_USERNAME=LOGIN-cc-us
PROXY_PASSWORD=PASSWORDRestart the stack (docker compose down, then docker compose up -d) so the API and the browser service start with the new values. The template comments the username and password lines out for proxies without a login; ours always needs one, so keep both.
Rotating or sticky
With a login such as LOGIN-cc-us, each connection leaves from a new US IP, which suits crawling many pages of many sites. For a site where a crawl has to stay on one IP, add a session key: LOGIN-cc-us-sid-crawl01-time-1800 keeps one exit for 30 minutes (6 hours at most). The whole instance shares that login, so every job then leaves from the same IP.
Drop the country key (LOGIN) to take exits in any country, or change it (LOGIN-cc-de) and restart to move the instance to another country.
SOCKS5
Firecrawl’s browser service runs Chromium through Playwright, which refuses SOCKS5 with a login. Use the HTTP port 7000 for every service.
Check it
Scrape https://api.ipify.org through your instance and read the result: it should be an exit IP, not your server’s. IP checker shows the country and network of that IP.
If it does not work
| What you see | What to check |
|---|---|
| Scrapes fail after the change | The login or the password, or no traffic left on the product (402). curl with the same login prints the gateway’s answer. |
| The result shows your server’s IP | The containers still run with the old .env: restart the stack. |
| Pages of one site come back blocked | A rotating login can look like many visitors to that site. Try a sticky key, or slow the crawl down. |
Questions
Not with these variables: they set one proxy for the whole instance. Run a second instance with another login, or fetch those pages with your own code, for example with Crawl4AI.
No. The variables belong to a self-hosted instance, the one you run yourself.