Skip to main content

migration / Checked 2026-10-18

Move long-running browser Functions from Steel to Notte

A migration plan for long-running browser Functions: retain the business check, replace Steel's provider-owned resources, and prove the Notte workflow before cutover.

Decide whether this move earns its cost

A deployed handler performs browser work longer than the caller is willing to wait. The execution must have an observable outcome even after the request ends.

Steel's managed and self-hosted choices make infrastructure ownership a concrete deciding factor. Keep that option on the shortlist when operating the browser is intentional; evaluate Notte when owning fewer execution and authentication components is the actual goal.

Notte is worth evaluating when a Python handler's run history can replace ad hoc background-job tracking. Confirm the actual runtime and cancellation contract rather than borrowing browser session limits.

An existing worker or provider execution service may fit a long task better when its runtime and cancellation behavior are already proven. A successful browser connection is not a runtime entitlement.

Keep the result contract stable

A safe delayed handler, a short caller wait budget and a stored execution reference.

The application can determine the execution's terminal state and browser ownership after the caller stops waiting, then retrieve or cancel work through the documented path.

Preserve the evidence contract across the move: Caller timeout, execution reference, handler terminal state, cancellation result and browser release.

Map the workflow's resources

RequirementSteelNotte
Code executionExternal workerPython Functions
Scheduled executionExternal schedulerNative
Recordings / replayDocumentedDocumented

Sources: docs.steel.dev [1] ↗docs.steel.dev [2] ↗docs.notte.cc [3] ↗docs.notte.cc [4] ↗

Where the code runs

Steel: The reviewed browser documentation establishes sessions, profiles and credentials. It does not establish a hosted Python handler with deployment and invocation APIs. An external worker can control a Steel browser; price and operate that worker separately.

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.

Sources: docs.steel.dev [1] ↗docs.notte.cc [2] ↗

Deployment ownership

Steel: Compare managed Steel with the documented self-hosted path explicitly. Self-hosting gives the team infrastructure responsibility that a hosted browser subscription would otherwise cover.

Notte: Use a remote browser from your existing process or deploy the browser automation as a Notte Function. These choices put lifecycle, scheduling and monitoring responsibilities in different places.

Sources: docs.steel.dev [1] ↗docs.notte.cc [2] ↗

Pricing model

Steel: Hosted browser plans and a self-hosted browser have different cost structures. For self-hosting, include the worker, compute, networking, storage and operations rather than assigning the open-source code a zero operating cost.

Notte: The public usage rates are $0.01 per browser-hour, $0.05 per Function-hour and $10 per GB of managed proxy traffic. Paid plans return their price as usage credits. Price the relevant meters and plan eligibility for the workload.

Sources: docs.steel.dev [1] ↗notte.cc [2] ↗

Cut over one responsibility at a time

For this Steel-to-Notte migration, keep the business assertion outside provider resource management. Source profile, connection, file and execution IDs are not valid target IDs; use a documented transfer path or establish the corresponding resource in Notte.

Retain source job history and a rollback route until the target pilot passes. A migration checklist is not evidence that saved state or runtime behavior is automatically portable.

  • Record source caller, handler and browser timeouts as separate budgets.
  • Port the handler without changing the business completion check.
  • Prove post-timeout status lookup, cancellation and browser cleanup before moving long jobs.

Sources: docs.steel.dev [1] ↗docs.notte.cc [2] ↗docs.notte.cc [3] ↗

A proposed acceptance case

This is an evaluation design, not a claim that this workflow has been executed successfully on every provider.

Part of the testRequired observation
Controlled inputA safe delayed handler, a short caller wait budget and a stored execution reference.
Pass conditionThe application can determine the execution's terminal state and browser ownership after the caller stops waiting, then retrieve or cancel work through the documented path.
Negative caseInterrupt the caller and request cancellation in a controlled test. Inspect both the handler outcome and whether its browser remains allocated.
Evidence to retainCaller timeout, execution reference, handler terminal state, cancellation result and browser release.

A reason to delay or keep the source

Interrupt the caller and request cancellation in a controlled test. Inspect both the handler outcome and whether its browser remains allocated.

Steel: Hosted and self-hosted Steel are different operating choices. A cloud feature's documentation does not establish identical packaging, retention or operational support in a self-hosted installation.

Delay cutover when the target's selected plan, required artifact path or supported website-account flow is not established. Resolve that gap before moving real jobs, even if the rate comparison favors the target.

Sources: docs.steel.dev [1] ↗notte.cc [2] ↗

Keep comparing

Put the workflow to the test.

Try Notte Read the quickstart ↗