Screenshot APIs all send your page content to their servers — built a self-hosted one, looking for feedback
Ran into this building an agent that needed to look at web pages. Every screenshot API is SaaS, which is fine, until you notice the page-understanding part runs on their model. So the text of every page you capture gets sent to a third party and processed there. For a public marketing page, who cares. For anything behind a login with real data on it, that's a problem I couldn't get around. Looked…
In the pursuit of building an agent that could analyze web pages, the author discovered that every screenshot API is a SaaS service. While this is acceptable for capturing marketing pages, it raises concerns when dealing with pages protected by login credentials containing sensitive information. The author sought a self-hosted solution, but found existing options like Urlbox and ScreenshotOne, which still send the captured content to a vendor's cloud servers.
Unable to locate a self-hosted screenshot API, the author decided to build their own. The extraction component of the tool can be integrated with any model, including Ollama running locally, ensuring that the page text remains within the machine where it was rendered. The software is released under the AGPL-3.0 license and can be deployed using Docker Compose.
However, the author acknowledges certain limitations of their creation. There is no security audit conducted, the extraction quality may not be exceptional, and the tool currently cannot capture private IP hosts due to an SSRF guard that blocks RFC1918 addresses without an opt-out (a feature the author plans to implement next).
The author also conducted a benchmark against two hosted APIs, but found that latency margins varied across runs, so they do not claim superior speed. All relevant data and caveats are available in the GitHub repository (github.com/route1-ai/shotbase).
The author is uncertain about the practical utility of their self-hosted screenshot API. Is running it yourself and keeping page content within your network the most valuable aspect, or do users primarily seek screenshots, leading to an overcomplication of the project? The author is genuinely seeking feedback on this matter, having built the tool for their own needs and unsure if it would generalize to other use cases.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.
