Hyperbrowser

Named integrations and custom tool support are both documented. LangChain, LlamaIndex, Stagehand and AI function-calling guides give an agent Hyperbrowser's browser, scrape and crawl as tools; HyperAgent's custom actions let a developer add their own tools; and x402 lets agents pay for fetches and searches.

Sourcehyperbrowser.ai/docs/integrations/langchainread 2026-09-23

Hyperbrowser runs hosted agent tasks: one API call starts a HyperAgent, Browser Use or model-native computer-use task (Claude, OpenAI, Gemini, Meta, Grok, Jev) in a managed session, the agent works through its steps, and the customer polls status, reads results or stops it. The open source HyperAgent adds multi-page management within one agent. That is multi-step execution of a single agent's loop, and no workflow composition, step chaining or multi-agent execution of Hyperbrowser's own is documented; action caching replays one agent's steps.

Sourcehyperbrowser.ai/docs/agents/overviewread 2026-09-23

Fetch, search, scrape, crawl and AI extract endpoints turn public web pages into markdown, HTML or structured data for the customer's own pipeline. That is web data for the customer to use rather than a retrieval structure Hyperbrowser maintains over the customer's own corpus; no index, vector store or grounding layer of Hyperbrowser's own over customer content is documented.

Sourcehyperbrowser.ai/docs/web/overviewread 2026-09-23

The human intervention surface is Live View: every session carries a Live View URL, embeddable in the customer's application, through which a person can watch and interact with the running browser, and the docs name human-in-the-loop interactions among its uses. No approval gate, pause primitive or policy guardrail is documented, and when to hand control to a person is decided in the customer's code.

Sourcehyperbrowser.ai/docs/sessions/live-viewread 2026-09-23

A Security Commitments page (last updated 7 May 2026) names the controls: an information security program aligned with the SOC 2 Trust Services Criteria, independent third-party assessments, annual third-party penetration testing, encryption at rest and in transit, vulnerability scanning, logging and monitoring, incident response, and backup and recovery. The attestation is unclear, with the pricing page's Enterprise card listing 'HIPAA/SOC 2' while the security page says only 'aligned with'; the page's access controls (least privilege, quarterly access reviews, SSO where available) govern Hyperbrowser's own team, and no SSO, roles or permissions for customer accounts are documented.

Sourcehyperbrowser.ai/securityread 2026-09-23

Hyperbrowser records the session. Web recordings (rrweb) capture DOM changes, interactions and network requests for replay, and MP4 video recordings are available too; both are fetched through the API, and every session has a live view. Data retention runs by plan: 7 days on Free, 30 on Startup and Scale, and 180 or more on Enterprise. That lets a buyer replay what the agent did in the browser. No structured per-action log of the agent's steps, no audit log separate from the recording, and no export to a monitoring system are documented.

Sourcehyperbrowser.ai/docs/sessions/recordingsread 2026-09-23

Sandboxes carry state forward: a memory snapshot freezes a running sandbox's memory, processes and filesystem and starts new sandboxes from that state, and volumes persist data across sandboxes, so the agent's own code reads the restored state to carry on. Snapshot state is the machine's, restored whole, and volumes are storage the code reads; no memory layer with a stated scope and lifetime is documented. Browser profiles, the saved login loaded at session start, and HyperAgent action caching, a replay of recorded actions without LLM calls, are not agent memory.

Sourcehyperbrowser.ai/docs/sandboxes/snapshotsread 2026-09-23

Buyers select from a published region list: browser sessions deploy to us-central, us-east, us-west, europe-west or asia-south, chosen per session. The security page adds that production data is hosted in United States databases unless otherwise agreed in writing. No self-hosted option is documented.

Sourcehyperbrowser.ai/docs/sessions/regionsread 2026-09-23

Ready-made starting points come as a templates gallery of ready-to-run sample projects built on the Hyperbrowser APIs (a news sentiment aggregator, product review analyzer, shopping recommenders, API servers wrapping Hacker News, Product Hunt and Craigslist, and others) and an examples repository. These are projects a developer clones and adapts rather than a catalog of agents a buyer adopts.

Sourcehyperbrowser.ai/templatesread 2026-09-23

Hyperbrowser starts work when a caller creates a session, sandbox, scrape job or agent task through the API or SDKs; no event, schedule, webhook or inbound queue of its own starts agent work. The action caching page lists scheduled jobs as a use case for replays, but the schedule is the customer's.

Sourcehyperbrowser.ai/docs/api-reference/create-new-sessionread 2026-09-23

The customer chooses the model behind Hyperbrowser's hosted agents. Tasks start with a chosen agent: HyperAgent or Browser Use, or model-native computer-use agents from Anthropic (Claude), OpenAI (CUA), Google (Gemini), Meta, xAI (Grok) and Jev. The pricing page bills each by its own model's tokens or steps, and HyperAgent supports OpenAI, Anthropic and Google Gemini models as configured providers.

Sourcehyperbrowser.ai/docs/agents/overviewread 2026-09-23

A documented REST API makes the platform callable from outside (sessions, profiles, extensions, sandboxes, volumes, snapshots, fetch, search, scrape, crawl, extract and agent tasks, with a full API reference), alongside official Python (pip install hyperbrowser) and Node.js (@hyperbrowser/sdk) SDKs, the hx CLI for sandboxes, and an open source MCP server.

Sourcehyperbrowser.ai/docs/sdks/introductionread 2026-09-23

This is where agents run, not how they are scored: no fixtures, scoring, quality gates or comparison of agent runs are documented as Hyperbrowser features. Session recordings and live view are the debugging surface, and action caching replays recorded actions deterministically to cut cost, which is execution rather than evaluation.

Sourcehyperbrowser.ai/docs/introductionread 2026-09-23

Operating a real browser for agents is Hyperbrowser's product: cloud browser sessions an agent drives through Playwright, Puppeteer or Selenium, computer actions (click, type, scroll, screenshot), or hosted computer-use agents from Anthropic, OpenAI, Google, Meta, xAI and Jev, with stealth, proxies, CAPTCHA solving, profiles and extensions for production use.

Sourcehyperbrowser.ai/docs/introductionread 2026-09-23