Deploy and Host Firecrawl on Sealos

Firecrawl is a self-hosted web crawling and scraping API that turns websites into clean, LLM-ready data. This template deploys Firecrawl with PostgreSQL, Redis, RabbitMQ, a Playwright rendering service, and public HTTPS access on Sealos Cloud.

About Hosting Firecrawl

Firecrawl provides API endpoints for scraping pages, crawling sites, and extracting content for AI workflows. The API service coordinates jobs, Redis stores queue and rate-limit state, RabbitMQ brokers background work, PostgreSQL stores Firecrawl data, and Playwright handles browser-rendered pages.

This Sealos template follows the official self-hosted topology while using Kubernetes-native services and managed database resources. Authentication is disabled by default for self-hosted API usage, matching Firecrawl's local self-hosting guidance.

Common Use Cases

• LLM data ingestion: Convert webpages into clean markdown or structured data for AI pipelines.
• Research crawlers: Crawl sites for documentation, market research, and content analysis.
• Browser-rendered scraping: Use Playwright-backed rendering for JavaScript-heavy pages.
• Private scraping API: Run a controlled Firecrawl endpoint inside your own Sealos workspace.

Dependencies for Firecrawl Hosting

The Sealos template includes the Firecrawl API container, Playwright service, PostgreSQL 16, Redis 7, RabbitMQ, internal Services, HTTPS Ingress, and App resources.

Deployment Dependencies

• Firecrawl Website - Product overview
• Firecrawl Documentation - API and SDK reference
• Firecrawl Self-Hosting Guide - Official self-hosted runtime guidance
• Firecrawl GitHub Repository - Source code and issues

Implementation Details

Architecture Components:

• Firecrawl API: Public API service on port 3002.
• Playwright Service: Internal rendering service used for browser-based scraping.
• PostgreSQL Cluster: Stores Firecrawl application data and the NUQ queue schema.
• Redis StatefulSet: Stores queue and rate-limit state.
• RabbitMQ StatefulSet: Provides the AMQP broker required by Firecrawl's job harness.
• Ingress and App Entry: Exposes the Firecrawl API through the generated Sealos HTTPS URL.

Configuration:

• Optional openai_api_key, openai_base_url, and model_name inputs enable AI extraction features.
• BULL_AUTH_KEY is generated automatically for the queue admin path.
• PostgreSQL, Redis, RabbitMQ, and Playwright URLs are wired internally by the template.
• The PostgreSQL init step creates the NUQ schema required by Firecrawl's self-hosted queue workers.
• The Firecrawl API runs with NUQ_WORKER_COUNT=1 and a 1536Mi memory limit based on live Sealos validation.

License Information:

Firecrawl is licensed under the AGPL-3.0 License. This Sealos template provides deployment configuration for running Firecrawl on Sealos Cloud.

Why Deploy Firecrawl on Sealos?

Sealos is an AI-assisted Cloud Operating System built on Kubernetes that unifies deployment and operations. By deploying Firecrawl on Sealos, you get one-click deployment, automatic HTTPS, managed data services, persistent storage, resource controls, and Canvas-based updates.

Deployment Guide

1. Open the Firecrawl template and click Deploy Now.
2. Configure optional OpenAI-compatible model settings if you need AI extraction.
3. Wait for deployment to complete, typically 3-5 minutes while PostgreSQL, Redis, RabbitMQ, and Playwright become ready. After deployment, you will be redirected to the Canvas. For later changes, describe your requirements in the AI dialog, or click the relevant resource cards to modify settings.
4. Use the generated public URL as your Firecrawl API base URL.
5. Test API availability with a /v1/scrape or /v1/crawl request from your client. The default template disables API authentication for self-hosted use.

Configuration

After deployment, configure Firecrawl through:

• API clients: Point SDKs or HTTP clients at the generated URL.
• AI Dialog: Update environment variables such as model settings or concurrency values.
• Resource Cards: Adjust API, Playwright, RabbitMQ, Redis, or PostgreSQL resources from the Canvas.
• Queue Admin Path: Use the generated BULL_AUTH_KEY if you expose and inspect the Bull queue admin route.

Scaling

Start with the default single API and Playwright replica. For higher crawl volume, increase API CPU and memory first, then review Playwright, Redis, RabbitMQ, and PostgreSQL readiness and storage.

Troubleshooting

API requests time out

• Cause: The API may still be waiting for Redis, RabbitMQ, PostgreSQL, or Playwright.
• Solution: Check all workload logs and database readiness from the Canvas before changing API settings.

API logs show worker load warnings

• Cause: Firecrawl's bundled queue worker is close to the configured CPU limit during crawl bursts.
• Solution: Increase the Firecrawl API CPU limit from the Canvas before increasing worker concurrency.

Browser-rendered pages fail

• Cause: The Playwright service is unavailable or under-resourced.
• Solution: Check the Playwright Deployment logs and increase CPU or memory for browser-heavy workloads.

AI extraction fails

• Cause: OpenAI-compatible model credentials are missing or invalid.
• Solution: Set openai_api_key, openai_base_url, and model_name for your provider, then restart the API Deployment.

Additional Resources

• Firecrawl API Documentation
• Firecrawl Self-Hosting Guide
• Firecrawl GitHub Issues
• Sealos Documentation

License

This Sealos template is provided as deployment configuration for Sealos users. Firecrawl itself is licensed under the AGPL-3.0 License.