Maxun v0.0.47 self-hosted upgrade: what operators should verify


Maxun v0.0.47 is worth a serious look if you want to self-host a platform that turns web pages into structured data and now watches them for changes. It fits teams willing to run and upgrade a Docker stack. It fits poorly if you need long-term stable releases or a vendor-backed SLA. Confirm registration, licensing, and the upgrade path before it touches production. This is a source and documentation review, not a hands-on test.

key facts

Item Detail
Author / reviewer Xiang Peng (XP812)
Review method Source and documentation review of the GitHub repository, official release notes, and docs; no local install or runtime test
Review date 2026-09-27
Version / release snapshot v0.0.47, published 2026-09-16; default branch develop
License AGPL-3.0
Stars (snapshot) About 17,565 stars, 1,519 forks, 106 open issues
Language TypeScript
Verification level observed
Material unknowns No runtime, performance, or upgrade-tested results; current status of the open registration issues; commercial support terms; data-residency behavior

what changed in maxun v0.0.47

The maintainers published v0.0.47 on 2026-09-16 under the tag v0.0.47. Their release notes name monitoring as the headline change: Maxun can track a website over time and notify you when it changes, either for a whole site or a specific part of it. The same notes say monitoring works with Extract, Scrape, and Crawl robots, is reachable through the no-code dashboard, SDK, CLI, and API, and is opt-in, so existing robots are unaffected.

The release also adds JPG and PNG input for document extraction and parsing, with OCR referenced in the changelog, plus run comparison for Scrape robots, a fix to authenticate the sockets the backend opens to itself, centralized run queuing, a fix for empty runs caused by page navigation, and an Azerbaijani translation.

Read that list as an operating signal, not a feature sheet. Several entries are reliability and authentication fixes in the run pipeline, and those shape how a self-hosted install behaves under automation. If you already run an older 0.0.x tag, they are the practical reason to schedule an upgrade window.

who should consider maxun, and who should not

Consider it if your team needs structured data from web pages and would rather operate a platform than maintain one-off scrapers. The project covers extraction of structured data, page-to-Markdown capture and screenshots, whole-site crawling, web search with time filters, document parsing, and now change monitoring. The no-code dashboard, API, SDK, CLI, and MCP surface (per the repository README) mean a small team can let non-developers build jobs while developers wire results into other systems.

Do not treat it as turnkey. If you need a long-term stable release line, a vendor SLA, or an environment where nobody can own a Docker stack, the current 0.0.x cadence and self-managed operations are a mismatch. If you need certified data controls, or cannot accept AGPL-3.0 obligations, resolve that before you adopt it. The hosted app is a different product surface, and this review cannot speak to its terms.

the self-hosted deployment has real moving parts

The project documents self-hosting and Docker Compose setup through its self-host documentation. Operationally, Maxun is not a single container. A user-reported Compose stack in issue #323 (opened 2025-01-07) runs a database, Redis, MinIO object storage, a backend, and a frontend, with environment variables for the database, object storage, Redis, and JWT and encryption keys. That report predates v0.0.47 and references old image tags, so treat the specifics as historical. The shape of the stack, though, a database plus cache plus object storage plus browser automation, is what you are signing up to operate.

The browser layer is the usual infrastructure pressure point for this class of tool. The same issue thread shows Chromium-oriented settings, including a larger shared-memory size and disabled sandboxing; check those against your host rather than copying them blindly. On footprint, the maintainers state in the v0.0.43 release notes that the backend image shrank by 61% and the browser image by 42%, which reduces pull time and disk use for self-hosted installs.

registration failures are the first thing to test after install

Several open issues describe accounts that cannot be created after a Docker deployment. Issue #534 (opened 2025-04-11) reports a registration error after changing only the public URLs from localhost to a device IP; it was still open and updated 2026-07-05 at this snapshot. Issue #323 reports the same class of failure with a full Compose configuration. Issue #306 is a duplicate report of the registration page not responding.

These reports pre-date v0.0.47 and may already be resolved, but they are not confirmed closed here. First-run registration and the public URL and backend URL configuration are the highest-value things to validate on your own hostname before you hand Maxun to anyone else. Budget that as a deployment step, and capture backend logs during a failed attempt instead of assuming a silent success.

what a 0.0.x upgrade cadence means for maintenance

From the release history, the project shipped v0.0.43 on 2026-06-24, v0.0.44 on 2026-07-15, v0.0.45 on 2026-08-03, v0.0.46 on 2026-08-25, and v0.0.47 on 2026-09-16: roughly monthly, pre-1.0 releases. That cadence is normal for a fast-moving project, and it means upgrades are a recurring task, not a one-off. The maintainers publish an upgrade guide covering Docker Compose and non-Docker setups, which is the right place to start, but the operator-side work is still yours.

Before each upgrade, pin the image tags you are moving from and to, back up the database and object-storage volumes, rehearse on a staging copy, and read the release notes for data or configuration changes. Version numbers around 0.0.x signal that interfaces and behaviors can still shift between releases.

license and data handling to confirm before you commit

Maxun is licensed AGPL-3.0. For teams that modify the code and let users interact with it over a network, the network-copyleft terms generally require offering the corresponding source to those users. If you plan to embed Maxun inside a closed product, get legal advice before you build on it, not after. The README also frames fair use of open source and asks commercial users to contribute back or sponsor. That is a community expectation rather than a license term, though it bears on how the project is sustained.

On data, scraping and monitoring raise questions the code cannot answer: whether the sites you target permit automated collection, where extracted data and screenshots are stored, how long you retain them, and who can access object storage. Decide those before you point the tool at production targets.

evaluation questions and alternatives

  • Does registration work end to end on your real hostname, not localhost?
  • Can you host a headless browser with enough memory and shared memory, and have you sized disk for screenshots and crawl output?
  • Do you have tested backups and a restore rehearsal for the database and object storage?
  • Who owns the AGPL-3.0 question for your intended use?
  • How will you pin versions, and who runs the recurring upgrade and verifies runs afterward?

On alternatives, weigh categories rather than specific vendors: the project’s own hosted app, commercial scraping and monitoring services that trade cost for zero ops, or a narrow in-house script for a single, stable target. Maxun’s advantage is breadth, self-hosting, and a no-code builder; its cost is that you own deployment, upgrades, and incidents. Pick based on which of those two you can absorb.

verdict

Maxun v0.0.47 is a credible self-hosted option for teams that want web extraction, crawling, and change monitoring in one platform, and the monitoring addition is a meaningful capability for operations use cases like tracking a page that should not change. The conditions are firm: you must be comfortable running a multi-service Docker stack, you should validate registration and URL configuration on your own infrastructure first, and you need a real plan for recurring 0.0.x upgrades and AGPL-3.0 compliance. If those hold, a staged pilot is worth it. If they do not, this is not the low-maintenance tool you are looking for.

how i can help

I take on self-hosted deployment, upgrade, and ongoing maintenance work for teams running platforms like Maxun. For this exact problem I would start by reviewing your deployment: the Compose or install setup, host resources, current version and how it was installed, the public URL and backend URL configuration, and the logs from a failed registration or upgrade attempt. I need that evidence and access first, plus your backup and restore setup, before I can say what is actually failing. From there I can plan the upgrade path, harden the deployment, and take over release support as a part-time technical owner. Reach me through contact, or see the wider software development services.

sources


Request help