The three-way buying decision
This shortlist starts with Kernel and Tilion, then includes Notte as a third implementation path for bursty browser workers. A queue receives more browser jobs than the account can run at once. Workers must wait for capacity and recover slots after failed jobs.
Browser slots and session creation rate constrain the burst independently. Queue work before allocating a browser, and apply bounded retries to avoid a synchronized retry wave.
A proposed acceptance case
A queue receives more browser jobs than the account can run at once. Workers must wait for capacity and recover slots after failed jobs.
This is an evaluation design, not a claim that this workflow has been executed successfully on every provider.
| Part of the test | Required observation |
|---|---|
| Controlled input | A fixed burst of jobs, the account's documented concurrency entitlement and a bounded application queue. |
| Pass condition | The queue drains without exceeding the entitlement; every allocated browser is associated with one job and eventually released. |
| Negative case | Fail a job after browser allocation. Verify that the next queued job can acquire the recovered slot and that retries do not create untracked sessions. |
| Evidence to retain | Queue arrival and start times, allocated sessions, completion counts and recovered capacity. |
Documented mechanisms, side by side
These labels describe product mechanisms. The linked source and the acceptance case determine whether a mechanism meets this workflow's requirement.
| Requirement | Kernel | Tilion | Notte |
|---|---|---|---|
| Browser control | Documented | Arena integration | Documented |
| Saved browser state | Documented | Not established | Documented |
| Live view | Documented | Not established | Documented |
| Deployment choice | Check deployment offering | Not established | Enterprise options |
Sources: kernel.sh [1] ↗kernel.sh [2] ↗kernel.sh [3] ↗kernel.sh [4] ↗tilion.com [5] ↗docs.notte.cc [6] ↗docs.notte.cc [7] ↗docs.notte.cc [8] ↗docs.notte.cc [9] ↗
Capacity and limits
Kernel: Kernel documents concurrent browsers, browser creation rate and per-browser resources separately. Standby browsers can still consume concurrency. Pools change how browsers are acquired; they do not erase account limits.
Tilion: Arena's observed test concurrency is a test configuration, not an account entitlement. Obtain the production concurrency and session creation limits for your access tier before sizing a fleet.
Notte: The public plans list 5, 25 and 100 concurrent browser sessions, with custom limits on Enterprise. Session duration and concurrent capacity are separate plan limits; a larger technical API limit does not override the plan.
Browser control
Kernel: Kernel exposes browser creation and control APIs. Its documentation distinguishes on-demand browsers from configured browser pools, so compare the acquisition path your application will actually use.
Tilion: BrowserArena includes a Tilion session integration. The reviewed public Tilion homepage does not publish the full session API contract; obtain the supported connection and lifecycle instructions through its access flow.
Notte: Create an owned browser session, connect Playwright or Puppeteer over CDP, and stop the session when the work finishes. Browser lifecycle management remains necessary even when the control socket disconnects.
Deployment ownership
Kernel: Kernel Apps move the automation process next to the browser. Alternatively, an external process can control a browser through the browser API. Choose based on execution locality and deployment ownership.
Tilion: The public access path currently points to a waitlist and contact. A production integration needs an agreed API contract, service access and ownership of the browser lifecycle.
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.
Pricing model
Kernel: Use the plan and browser configuration in Kernel's pricing documentation. Managed networking, account allowances and app execution make a browser-hour comparison only one part of the decision.
Tilion: Arena provides an estimated usage rate for its test. The reviewed public homepage does not establish a self-serve production plan, allowances or overage terms. Obtain those terms before quoting a monthly bill.
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.
What should change the choice?
Notte is a candidate when its published concurrency tiers fit the queue and the team wants browser and Python execution resources in one platform. Capacity must be checked separately from the hourly rate.
Kernel: Kernel's browser-local Apps make execution placement part of the decision. Compare the repeated action path with the proposed Notte controller or Function deployment before concluding that a lifecycle or hourly-rate difference decides the workflow.
Tilion: The reviewed Tilion homepage offers a waitlist and contact route. Arena's tested integration is useful evidence about a run, but a purchasing or migration decision still needs a documented access, state and execution contract; those unknowns must remain visible beside Notte's public docs.
An existing pool or queue integration can outweigh a rate difference when it already handles the burst reliably. A single-session latency result does not establish fleet behavior for any provider.
Sources: kernel.sh [1] ↗kernel.sh [2] ↗kernel.sh [3] ↗kernel.sh [4] ↗tilion.com [5] ↗notte.cc [6] ↗docs.notte.cc [7] ↗docs.notte.cc [8] ↗
How to remove an option from the shortlist
For this workflow, require the following evidence from Kernel, Tilion and Notte: Queue arrival and start times, allocated sessions, completion counts and recovered capacity.
Remove a proposed configuration when its documented mechanism or a controlled pilot cannot meet the pass condition. Keep an undocumented feature as an unresolved question instead of inventing an equivalence. Do not select a provider solely because it appears in this shortlist.
- The queue drains without exceeding the entitlement; every allocated browser is associated with one job and eventually released.
- Fail a job after browser allocation. Verify that the next queued job can acquire the recovered slot and that retries do not create untracked sessions.
Make a bounded pilot
Measure time to drain a fixed burst, total attempted starts, task success and the bill for both successful and failed work.
Use the same authorized account, input and final-state check in each configuration. Treat a missing product contract as an open question; do not turn an undocumented capability into an unsupported claim or a successful test result.
- Keep admission control in place while introducing the target provider.
- Move one queue partition and record allocation failures separately from task failures.
- Compare drained jobs, leaked slots and cost per completed job before raising concurrency.