The language of agentic browser automation

A vector cache that matches a new, re-phrased request to a skill you already compiled — so a repeated task skips the LLM.

Re-running a compiled skill as a fixed sequence of indexed actions, with zero LLM calls, so the same goal produces the same path every time.

Turning a successful agent run into a reusable, replayable skill — a compact program of indexed actions you can call again without re-planning.

A compact, numerically-indexed map of a page’s interactive elements that an agent acts on by index instead of raw HTML.

A shared, sanitized library of compiled skills that lets a skill compiled once be safely reused across tenants, compounding the cache savings.

Pausing a run at a step that needs a person — an approval or an MFA prompt on an authorized flow — then resuming automatically once it’s cleared.

The infrastructure that turns an agent’s goal into deterministic, replayable browser action — the layer between your LLM and a real browser.

A page representation stripped to just the interactive, indexed elements under a token budget — so prompts stay small and cheap.

A Model Context Protocol endpoint that exposes Twin as agent tools (run, compile_skill, run_skill) to MCP clients like Cursor, Claude Desktop, Claude Code, and Cline.

Headless runs a browser with no visible UI; headful runs a real, rendered browser. The choice affects compatibility, observability, and how some sites behave.

RPA follows brittle, hand-recorded scripts; agentic automation has a model plan against the live page — then compiles the result so it can replay cheaply.

A durably-stored recording of a browser run, plus real-time live view — so every automated session is watchable and auditable after the fact.

Encrypted per-tenant storage for the secrets a run needs — logins and tokens — injected at execution time and never exposed to the model or other tenants.

The cost of one more execution of a workflow. On most browser infra it’s flat; on Twin it falls as cache hits replace LLM calls.