Browserbase by Browserbase

# Fix list: Browserbase

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/browserbase, the October 2026 research run, assessed 1 October 2026. Grade BB, 76.6 out of 100.

This is everything the published grade says the listing lacks, the biggest possible gain to the total first. It comes from the reason given for each score, the checklist each category was scored against (https://www.anchorterminal.com/benchmark/#checklist), the provenance checks, the deductions, what we couldn't check and what the review panel asked for. A fix counts at the next check, once it's public.

For a coding agent working on Browserbase: work through the items below in the product, its docs and its public pages. Each category gives the reason for its score, with the points each checklist item earned, and the checklist itself, so the gap is the items that earned less than their points. Change the product, not the wording, and keep a note of what you changed and where it's published.

## 1. Security & auth, 43 out of 100, up to 10 more on the total

Why it scored 43: A project API key in the `X-BB-API-Key` header, and we found no scopes, rotation guide or per-key permissions, so one key controls the project (10). The hosted MCP setup page shows the key in the URL as `?browserbaseApiKey=`, a secret in a query string (minus 10). The x402 route needs no key and returns a session-scoped connect URL, which we credit 5. No read-only mode for the MCP server, but each browser runs in its own VM on an isolated subnet, and Stagehand 3.7 added domain policies (8). Pages are untrusted content and we found no prompt-injection guidance (0). Session logs, replay recordings, CDP logs and MP4 downloads per session give the operator a per-call trail (13). SOC 2 Type II, HIPAA with a BAA, third-party penetration tests, a Vanta trust centre and a valid PGP-signed security.txt expiring 2027-06-01. No bug bounty found (17).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-security):

- 0 to 30, the credential model. 30 for OAuth 2.1 with scopes, or scoped and revocable keys with rotation. 20 for plain revocable API keys. 10 for one all-powerful key. 10 off when a secret can travel in a URL query string as a documented option.
- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.
- 0 to 15, prompt-injection posture where the tool returns untrusted content (documented mitigations or guidance). A tool that returns no untrusted content gets 10.
- 0 to 15, audit logs or per-call visibility for the operator.
- 0 to 20, a security programme. security.txt or a disclosure policy, a bug bounty, SOC 2 or ISO 27001, advisories handled in public.

Models are read for retention, whether API data trains models (and whether that's off by default), zero-retention options and certifications. Frameworks for telemetry defaults, approval hooks, guardrails and sandboxing.

## 2. Agent ergonomics, 75 out of 100, up to 4.1 more on the total

Why it scored 75: The hosted MCP server has six compact tools, and the API lists take `limit`, so context cost is low. Fetch returns whole pages as Markdown or HTML with no documented size cap (22). Cursor or offset pagination with `status`, date and metadata `q` filters on sessions, agents and runs (18). Error bodies are documented for Fetch and downloads only (12). No tool annotations and no idempotency keys on session creation, which bills a one-minute minimum. The x402 route refunds unpaid minutes on termination (8). Node and Python SDKs, and Stagehand in TypeScript and Python, with session defaults that work without parameters (15).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-ergonomics):

- 0 to 25, context cost. For MCP, the number and size of the tool definitions (25 for ten or fewer compact tools, 15 for 11 to 30, 5 for more than 30, plus up to 10 back for toolsets, dynamic loading or read-only subsets). For APIs, whether responses can be sized (field selection, limits, summaries).
- 20, pagination, filtering and output-size controls.
- 20, actionable, documented error responses, codes and messages an agent can recover from.
- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.
- 15, sensible defaults, few required parameters, and official SDKs in at least two languages.

Models are read for tool use, structured output, prompt caching, context length, batch and SDKs. Frameworks for how much code and how many defaults a tool-calling agent with MCP needs.

## 3. Schema & documentation, 80 out of 100, up to 3.3 more on the total

Why it scored 80: OpenAPI 3.0.0 at docs.browserbase.com with 22 path groups and the `X-BB-API-Key` header scheme (25). llms.txt and Markdown twins of the docs (10). The API reference reads well, but the hosted MCP server's six tools carry one-line descriptions such as "Perform an action on the page" with no when-not-to guidance (10). The OpenAPI document types inputs with ranges such as `timeout` 60 to 21600 seconds, while the MCP tools take one free-text string each (11). Error schemas exist for `/v1/fetch` (400, 402, 403, 429, 502 to 504) and recording downloads, and most other endpoints have none. Code samples are plentiful (9). `/v1` versioning and a dated public changelog with 12 entries since 13 July (15).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-schema):

APIs and MCP servers.

- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).
- 10, llms.txt or Markdown docs served for agents.
- 0 to 20, descriptions that say what a tool is for, when to use it and when not to, read from the tool definitions in the source or the API reference.
- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.
- 0 to 15, examples and documented error responses.
- 15, versioning and a public changelog.

Models are read from the API reference, the OpenAPI file, llms.txt, the structured-output and tool-use docs and the model cards. Frameworks from docs a model can follow, typed interfaces, examples and the API reference.

## 4. Transparency & trust, 75 out of 100, up to 2.2 more on the total

Made of editorial 49, provenance 100.

Why it scored 75: The platform is closed with published terms, and Stagehand is MIT (20). The privacy policy (last updated 1 June 2024) keeps recordings 30 days, while the pricing page gives Free plans 7 days. `recordSession` and `logSession` can be turned off per session for zero retention. No DPA or subprocessor list in the policy (14). No deprecation policy. The open-source MCP repository was archived on 20 July 2026 with a notice, but the MCP setup page still describes self-hosting it without saying so (5). Regions documented as US West, US East, EU (Germany) and Singapore. Subprocessors only through the Vanta trust centre, which didn't render for us (10).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-transparency):

- 0 to 30, source availability and licence clarity. 30 for open source under an OSI licence, 15 for closed with clear terms, 0 for unclear terms.
- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).
- 0 to 20, a deprecation policy or notices with dates.
- 0 to 20, telemetry disclosed with an opt-out (local software), or subprocessors and data locations disclosed (hosted).

The other half of Transparency and trust is the provenance score, computed from checked facts (below). The category score is the mean of the two.

## 5. Reliability, 90 out of 100, up to 2 more on the total

Why it scored 90: status.browserbase.com runs on incident.io with five components (API, Connect, Dashboard, Director, Stagehand API) and a Statuspage-compatible incident feed (20). The feed lists 26 incidents from December 2024 to 26 May 2026, the last a 49-minute critical dashboard login outage, and nothing in the 90 days to this check (30). Concurrency and session-creation limits published per plan, 3 and 5 a minute on Free up to 250-plus and 150-plus on Scale (15). 429 with `retry-after` and `x-ratelimit-*` headers, and a documented retry helper with exponential backoff for session creation (15). No SLA on the pricing page or in anything else we read (0). The sessions API is GA (10).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-reliability):

Hosted APIs, MCP servers, models and platforms.

- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).
- 0 to 30, the incident record for the last 90 days on that page. 30 for a clean record or trivial incidents only, 20 for minor incidents only, 10 for one major outage (an hour or more of a core API down, or errors across the board), 0 for several. 5 when there's no history we could read, and the note says so.
- 15, rate limits documented with numbers.
- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.
- 10, an SLA published for any paid tier.
- 10, the surface agents use is generally available, not beta or preview.

Local packages, SDKs, frameworks and stdio MCP servers.

- 20, installs from an official package with supported runtimes stated.
- 25, a public CI and test suite, passing on the default branch.
- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).
- 15, semver discipline and breaking changes called out in a changelog.
- 15, version 1.0 or later, or declared stable.

Protocols are read from their reference implementations, the public facilitators or servers, spec stability and test vectors.

## 6. Payments & pricing, 90 out of 100, up to 1.3 more on the total

Why it scored 90: x402 on x402.browserbase.com for browser sessions (create, status, extend, terminate) at $0.12 an hour in USDC on Base, with a refund of unused minutes on termination. Fetch, Search and the main api.browserbase.com host aren't covered. Stripe's 18 March 2026 MPP post names Browserbase as selling per-session browsers, so partial (30). Per-hour and per-1,000-call prices published without login (20). Free plan with 1 browser-hour, 3 concurrent browsers and 1,000 Fetch and Search calls. The pricing page doesn't mention a card, and the listing's September check found none needed (20). An agent with a wallet gets a browser with no account (20).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-payments):

The published rubric, also on the [x402 page](https://www.anchorterminal.com/x402/).

- 40, a machine payment protocol (x402, MPP or L402) on the tool's own endpoints. 10 to 30 when it covers only some endpoints or only goes through a third party, and the note says which.
- 20, per-call or per-unit pricing published without a login. 10 for public plan-only pricing, 0 for "contact sales" or prices behind a login.
- 20, a free tier or trial that doesn't need a card.
- 20, autonomous onboarding, meaning an agent can get access without a person signing up in a browser (keyless use, x402, a programmatic key API).

Payment platforms and agent wallets rarely charge for their own API over a machine protocol, so the first line has steps for them, and the highest one that applies counts. 40 when x402, MPP or L402 runs on all their own endpoints, 30 when it runs on part of their own API, 25 when their merchants can accept one, 20 for running a facilitator, 15 for paying as a buyer, and 0 when the only protocol is their own. Merchant acceptance sits above a facilitator because the platform's own customers can charge agents through it, while a facilitator settles for sellers who wire up the protocol themselves. The counter-argument (a facilitator does more for the protocol as a whole) has a point. Each note says which step applied.

Open-source software you run yourself is scored on its hosted or paid option if it has one. A free, self-hosted package with nothing to buy gets 20, 20 and 20 for the last three lines, and 0 to 40 for the first only if it ships a payment protocol.

## 7. Maintenance & community, 92 out of 100, up to 0.7 more on the total

Why it scored 92: Changelog entry on 30 September 2026 and Stagehand 4.1.0 on 9 September (30). Twelve dated changelog entries since 13 July and five Stagehand 4.x releases since 9 August (20). Stagehand merges commits daily, including mirrored external pull requests, but we didn't read open-issue reply times (17). Current official SDKs in Node and Python. The only Browserbase-owned MCP registry entry is the archived stdio server at 2.1.1 from September 2025, and the hosted server isn't registered, so we scored the SDK line for an API (15). Stagehand CI runs on every merge (10).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-maintenance):

- 0 to 30, time since the last release, or the last published model or API change for a closed service. 30 within 30 days, 20 within 90, 10 within 180, 0 older.
- 20, at least three releases or dated changelog entries in the last 90 days.
- 0 to 25, responsiveness. Issues and pull requests answered on GitHub (the open issues and how recent the replies are). For closed services, a public changelog and a support or community channel that answers, 0 to 15.
- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).
- 10, package health, current dependencies and CI.

Models are read for deprecation notice periods and model churn rather than release counts.

## What we couldn't check

What we couldn't read counted as absent. Publishing it on a page a plain HTTP fetch can read (not only in a browser) lets the next check count it.

- unchecked: the subprocessor list, held in the Vanta trust centre at trust.browserbase.com, which didn't render for our reader
- unchecked: whether the free plan asks for a card. The pricing page doesn't say and we relied on the listing's September check
- unchecked: whether API keys can be scoped, rotated or revoked individually
- The listing said Browserbase had no x402. Its docs now document x402 sessions at x402.browserbase.com, so we patched the x402 block
- Whether the incident feed is complete after an apparent move from Statuspage to incident.io. It lists nothing after 26 May 2026

## Weaknesses

- The hosted MCP setup page passes the API key as `?browserbaseApiKey=` in the URL
- No scopes or rotation guide for API keys, and no SLA published
- Hosted MCP tools have one-line descriptions and no annotations
- Privacy policy last updated 1 June 2024 says 30-day recordings, the pricing page says 7 days on Free
- The open-source MCP repository was archived on 20 July 2026 but the setup page still describes self-hosting it

## What costs an agent a turn today

The notes we give agents before they call it. Each one is a workaround an agent shouldn't need.

- Use x402.browserbase.com when you have a wallet and no key. Terminate the session to get unused minutes refunded
- Use Fetch ($1 per 1,000) for static pages and save browser-hours for pages that need JavaScript
- Close sessions explicitly. Each session bills at least one minute, and idle ones keep billing
- Set `recordSession` and `logSession` to false when pages hold personal data
- On 429, wait `retry-after` seconds before creating another session

## What the review panel asked for

- x402 beyond sessions
- scoped API keys
- a registry entry for the hosted server
- a written deprecation policy
- Idempotent session creation
- Write when-not-to-use text
- Error schemas on every endpoint
- name Search's sources
- document a Fetch size cap
- Session creation idempotency
- Publish an SLA
- Idempotent session create
- Header-only MCP auth
- header-only MCP auth
- scoped, rotatable API keys

## When it's done

Send what changed and where it's published as a dispute (https://www.anchorterminal.com/builders/#disputes, or `POST https://www.anchorterminal.com/api/v1/contact` with `"kind": "dispute"`). Disputes are answered in public, and the listing is checked again by the same checklist. Paying for an audit or a listing claim changes nothing here.