Can you explain what happened after a run fails?
An active live view helps someone inspect or take over a browser. A recording explains a finished run. Application logs, network events and Function outputs add different context; none alone proves that the requested operation succeeded.
A screenshot showing a confirmation page can support a result, while a saved job identifier lets the application verify it later. Capture both when the workflow creates a persistent object.
Compare the evidence from a run
Notte: Live view shows an active browser and can be embedded in an application. Recordings let you review a finished session. Function runs have their own history and results, so keep the run ID alongside the browser session ID.
Kernel: Live browser URLs support an embedded view and a read-only option. Replays and browser telemetry are separate observability surfaces; app invocation logs add execution context.
Sources: docs.notte.cc [1] ↗kernel.sh [2] ↗
Compare the execution model
Notte: Functions deploy Python code with a run() entry point. Invoke them over HTTP or with an SDK, inspect run history, and roll back code by redeploying an earlier version as the latest version.
Kernel: Deploy a TypeScript or Python App and expose Actions that callers invoke. Kernel co-locates code with the browser environments the app creates. An Action invocation is the execution unit rather than a bare CDP connection.
Sources: docs.notte.cc [1] ↗kernel.sh [2] ↗
Compare the login path
Notte: Vaults substitute credential placeholders in supported browser actions outside the model context. Managed Auth checks requested saved connections before automation. Personas supply an inbox and optional phone number for verification workflows.
Kernel: Managed Auth handles website login and eligible connection recovery. Vault fill is a separate flow for applications that control navigation and form submission; the reviewed vault documentation labels that surface as preview.
Sources: docs.notte.cc [1] ↗kernel.sh [2] ↗
How to evaluate this choice
Choose observability around how the team diagnoses failures. A visually rich replay can be useful, but an error message and the exact account, input and code version often determine whether the failure can be reproduced.
- Keep a correlation between the application job, browser session and Function run.
- Enable the required recording mode before the run starts.
- Check retention, export and access behavior for the chosen plan.
- Verify whether a viewer is interactive or read-only before handing control to a user.
- Record the final application state as well as a video or log.
Requirements side by side
| Requirement | Notte | Kernel |
|---|---|---|
| Browser control | Documented | Documented |
| Saved browser state | Documented | Documented |
| Code execution | Python Functions | TypeScript / Python Apps |
| Deployment choice | Enterprise options | Check deployment offering |
Sources: docs.notte.cc [1] ↗docs.notte.cc [2] ↗docs.notte.cc [3] ↗docs.notte.cc [4] ↗docs.notte.cc [5] ↗docs.notte.cc [6] ↗kernel.sh [7] ↗kernel.sh [8] ↗kernel.sh [9] ↗kernel.sh [10] ↗kernel.sh [11] ↗kernel.sh [12] ↗
Which approach fits?
Choose Notte when your browser workflow needs saved logins, credential fill or scheduled Python code alongside the browser.
Notte tradeoff: A saved profile restores browser state; it does not prove that a website still considers the user signed in. Use Managed Auth when the workflow needs connection readiness before it starts.
Choose Kernel when your code makes many browser calls and you want to deploy it beside the browser, or you need Kernel's managed sign-in flow.
Kernel tradeoff: Browser-local code changes the deployment model. Check app action behavior, account limits and persistent-state ownership before moving an external worker into it.
Sources: docs.notte.cc [1] ↗kernel.sh [2] ↗
A useful acceptance check
Write a pass/fail result for the real task before running the evaluation. Record the account, configuration and code version next to that result. Repeat failures need an explanation, not an automatically different browser configuration.