What is open, and under which licence
| Part | Licence | What it does |
|---|---|---|
| Crawling engine | AGPL-3.0 | Fetching, rendering and HTML-to-markdown. A fork of the open-source Firecrawl engine. |
| Pipeline and API | Apache-2.0 | Jobs, knowledge bases, document parsing, OCR, chunking, embeddings. |
| SDKs, CLI, MCP server, n8n node | Apache-2.0 | Every client surface. |
| Search service | Hosted only | Keyword and hybrid search over a knowledge base, on paid plans. |
Self-hosting, honestly
The stack ships as a Docker Compose file for a single host and a Swarm bundle for a cluster, behind one nginx front door. It checks five dependencies at boot: PostgreSQL, Redis, RabbitMQ, MinIO and Elasticsearch.
It is not small: around 23 services and about 60 GB of free disk before it starts. That is the honest cost of running crawling, rendering, OCR and indexing yourself, and it is why most teams start on the hosted API.
- Self-host when data must stay in your network, you need to modify the pipeline, or per-page pricing does not fit your volume.
- Use the hosted API when you would rather not operate a browser pool, a queue and a search cluster.
Compared with other open-source scrapers
| SarvCrawl | Firecrawl (self-hosted) | Crawl4AI | |
|---|---|---|---|
| Licence | Apache-2.0 + AGPL-3.0 engine | AGPL-3.0 | Open source |
| Output | Markdown, HTML, JSON, links | Markdown, HTML, JSON, links | LLM-ready markdown |
| PDF / Office parsing + OCR | Yes | PDF parsing | PDF support |
| Chunking and embeddings | Built in | You build it | You build it |
| Shape | Full stack with a console | API service | Python library |
Questions
Is SarvCrawl open source?
The pipeline, SDKs, CLI, MCP server and console are Apache-2.0, and the crawling engine is AGPL-3.0 as a fork of Firecrawl. The search service is hosted and is not part of the open-source release.
Can I use it commercially?
Yes, under the terms of each licence. Apache-2.0 is permissive. The AGPL-3.0 engine requires you to offer its source to users if you modify it and run it as a network service.
What do I need to run it?
Docker, and a host with room for about 23 services and 60 GB of free disk. The Compose file brings up PostgreSQL, Redis, RabbitMQ, MinIO and Elasticsearch alongside the app.
Is the hosted version different?
It is the same pipeline, plus hosted keyword and hybrid search on paid plans, and nobody on your team has to operate it.