authentik
The first patch on 2026.8 is twenty-odd fixes and no new surface.
A side-by-side editorial comparison of Retool and TestingBot — release velocity, themes, recent moves, and the top alternatives to consider.
Retool posts one-line changelog entries, and its agent work is the only visible thread
Retool publishes very short changelog entries — most under 120 characters, several under 60 — so each release is a headline with a sentence attached. The recent run covers message queueing, branch management gaining rename and copy, performance work on the app building agent, a BigID resource deprecation, and periodic self-hosted stable tags. Nothing here carries enough text to judge scope, only direction.
A browser grid that spent the summer becoming a team product, and is now opening itself to agents.
TestingBot runs a cross-browser and real-device cloud, and the last three months have gone almost entirely into what surrounds a test run rather than the run itself: custom roles enforced identically in dashboard and API, service accounts so pipelines stop borrowing a person's credentials, enforced two-factor authentication, comments with @mentions on failed runs, one-click filing into GitHub, Jira, Trello, Asana or Azure DevOps, and a notification inbox tying it together. The grid work continues underneath — WebRTC streaming for iOS simulators, a rebuilt web inspector, new devices and browser versions on the usual cadence.
Retool publishes very short changelog entries — most under 120 characters, several under 60 — so each release is a headline with a sentence attached. The recent run covers message queueing, branch management gaining rename and copy, performance work on the app building agent, a BigID resource deprecation, and periodic self-hosted stable tags. Nothing here carries enough text to judge scope, only direction.
Two threads are legible from the headlines. The app building agent keeps getting attention, which is where Retool is putting its AI bet, and the branch-based development workflow keeps accruing the controls a team needs — renaming and copying branch names is small, but it is the kind of gap that only matters once people are actually working in branches. Deprecations like the BigID resource type suggest a periodic prune of connectors that did not find usage.
Branch controls have now been extended twice without a governance layer appearing, so expect permissions or review on branch operations next, since that is what teams ask for once branch-based development is in real use.
TestingBot runs a cross-browser and real-device cloud, and the last three months have gone almost entirely into what surrounds a test run rather than the run itself: custom roles enforced identically in dashboard and API, service accounts so pipelines stop borrowing a person's credentials, enforced two-factor authentication, comments with @mentions on failed runs, one-click filing into GitHub, Jira, Trello, Asana or Azure DevOps, and a notification inbox tying it together. The grid work continues underneath — WebRTC streaming for iOS simulators, a rebuilt web inspector, new devices and browser versions on the usual cadence.
Two directions are visible. The platform is being made safe for organisations rather than individuals, which is what roles, service accounts, 2FA and audit-friendly reporting add up to. Alongside it, the grid is being made reachable by software instead of people: an Insights API exposing pass rates and flakiness, embeddable build reports for CI dashboards, GitHub checks on every pull request, and now Vibium's CLI and MCP server pointed straight at the grid over WebDriver BiDi. The AI work so far reads results — failure analysis, documentation questions — while Vibium is the first path where something other than a human-written script drives the browser.
Expect the agent-facing path to be developed further, since supporting Vibium's MCP server without extra configuration is a small step from offering a first-party one, and expect the device and browser catalogue to keep rotating on its established cadence.
Other Infra & APIs products tracked by Sparkpulse, ranked by recent ship velocity. Each card links to a full editorial trajectory and lets you pivot into a head-to-head comparison with either Retool or TestingBot.
The first patch on 2026.8 is twenty-odd fixes and no new surface.
Alert rate limits stop being per-source and start being per-tenant, by query.
The essay series keeps running ahead of the products it describes.
Both lines patch on the same day, and the reverted JSONata upgrade finally sticks.
The 3.x line finally ships function: authenticated secret writes and render patches.
Folder-level RBAC lands, and approvals finally get a webhook to talk to.
See all Retool alternatives → · See all TestingBot alternatives →
Latest ship moves from both products, interleaved chronologically. ⚡ = editorial spark.
They serve adjacent needs but don't currently overlap on shipped themes. Retool and TestingBot are shipping at a similar cadence (velocity 5.0 vs 5.0, both within Sparkpulse's "active" band). See the at-a-glance table above for a side-by-side breakdown of velocity, recent sparks, and editorial themes.
Sparkpulse doesn't pick a winner — we score release velocity, not feature parity. Retool and TestingBot are shipping at a similar cadence (velocity 5.0 vs 5.0, both within Sparkpulse's "active" band). For your specific use case, the alternatives sections above list other Infra & APIs products to evaluate alongside.
Top Retool alternatives in Infra & APIs are ranked by recent ship velocity. Browse the "Retool alternatives" section above for the current picks, or visit /alternatives/retool for the full list with editorial commentary on each.
Top TestingBot alternatives in Infra & APIs are ranked by recent ship velocity. Browse the "TestingBot alternatives" section above for the current picks, or visit /alternatives/testingbot for the full list with editorial commentary on each.