Back to vendors
S

Scrapybara

Also known as: Scrapybara Act SDK

Visit site
Entry priceFrom $29/mo · free tierFull pricing detail

Computer-use and browser agent infrastructure with the Act SDK for web-based agentic automation tasks.

Scrapybara is infrastructure that gives AI agents a full computer to work in. While many models are capable reasoners, they are effectively disembodied: they cannot operate the large share of digital work that lives in graphical interfaces and desktop software rather than clean APIs. Computer-use agents address this by controlling a real machine the way a person does, with a screen, keyboard, mouse, and filesystem, and Scrapybara provides the hosted virtual desktops those agents run on, so developers can focus on agent logic instead of provisioning and scaling machines.

With a single API call, a developer can spin up an isolated virtual desktop in seconds, available as both Ubuntu Linux and Windows environments, each with a browser, a filesystem, and code execution built in. The platform is designed for scale: hundreds of instances can start almost instantly, and Scrapybara handles the autoscaling, authentication, and system environment so fleets of agents can be deployed to production without infrastructure overhead.

A defining piece is its unified interface for computer-use models. Through an Act SDK, a developer connects an instance to whichever computer-use model they prefer, such as OpenAI's Computer-Using Agent or Anthropic's Claude Computer Use, and drives it with tools for the computer, the shell, and file editing, getting structured outputs and multi-turn conversations from one consistent API rather than wiring each model up separately. Low-level controls expose the browser, filesystem, environment variables, and code sandboxes directly.

Several features make these agents practical in the real world. Authenticated browser sessions can be saved and reused across instances so an agent does not have to log in repeatedly, instances can be paused and resumed to preserve work, and an interactive stream lets a developer watch the agent act and step in when needed. Self-hosted deployment is available for teams with security or compliance requirements. In practice, Scrapybara targets work that browser-only tools cannot reach, such as automating legacy desktop applications, operating GUI-only enterprise software, running parallel research across many machines, and executing end-to-end UI tests, giving agents the equivalent of a remote worker's computer.

Vendor details

Canonical URL

https://scrapybara.com/

Category

Browser / computer-use agent

Company status

unclear

Use cases & customers

Target customers

developers

Deployment options

SaaSAPI

In practice

Your agent needs to operate a legacy desktop application or GUI-only enterprise software that has no API. Scrapybara spins up a full Ubuntu or Windows virtual desktop where a computer-use agent can click, type, and run software like a person.

You want to run computer-use agents at scale without provisioning machines yourself. Scrapybara starts hundreds of isolated desktop instances on demand and handles autoscaling, authentication, and system environments for you.

You're building on a computer-use model and don't want to wire up infrastructure for each one. Scrapybara's unified Act SDK connects an instance to models like Claude Computer Use or OpenAI's CUA through one consistent API.

Agentic Index coverage score

9.0 / 14 capabilities · 64%

Integrations & Tool CallingAgentic AI Index Expansion Evidence Report Full
Workflow OrchestrationAgentic AI Index Expansion Evidence Report Full
Knowledge Grounding & RAGAgentic AI Index Expansion Evidence Report Unable to verify
Human Oversight & GuardrailsAgentic AI Index Expansion Evidence Report Partial
Security, Identity & GovernanceAgentic AI Index Expansion Evidence Report Partial
Observability & AuditabilityAgentic AI Index Expansion Evidence Report Partial
Memory & State PersistenceAgentic AI Index Expansion Evidence Report Partial
Deployment & Data ResidencyAgentic AI Index Expansion Evidence Report Unable to verify
Prebuilt Agents, Templates & PacksAgentic AI Index Expansion Evidence Report Full
Triggers & Channel CoverageAgentic AI Index Expansion Evidence Report Partial
Model Flexibility & RoutingAgentic AI Index Expansion Evidence Report Full
APIs, SDKs & MCP ExtensibilityAgentic AI Index Expansion Evidence Report Full
Testing, Debugging & OptimizationAgentic AI Index Expansion Evidence Report Partial
Browser & Computer UseAgentic AI Index Expansion Evidence Report Full

The Agentic Index coverage score grades every vendor Full, Partial or Unable to verify against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded

Pricing

From $29/mo · free tier

usage

Free tier

Included quota

Free: 10 compute hours, 100 agent credits, 5 concurrent instances. Pro (~$29/mo): 100 compute hours, 500 credits, 25 concurrent. Business (~$99/mo): 500 hours, 2,500 credits, 100 concurrent. Enterprise: custom. Top-ups $0.04/credit; BYO provider key offsets model cost.

What is public

Scrapybara (scrapybara.com - virtual-desktop / computer-use infrastructure for AI agents: spin up Ubuntu/browser desktop instances at scale with browser, code-execution, and filesystem capabilities; Act SDK + ComputerTool/BashTool/EditTool; handles orchestration/scaling) publishes usage-based pricing with tiers. Free: 10 compute hours, 100 agent credits, 5 concurrent instances. Pro ~$29/mo: 100 compute hours, 500 credits, 25 concurrent. Business ~$99/mo: 500 hours, 2,500 credits, 100 concurrent. Enterprise: custom. Billing = compute hours (instance runtime) + agent credits (1 credit per step = 1 assistant message with tool calls via Act SDK/playground). Credit top-ups $0.04 each, or bring your own provider API key.

Billing mechanics

Usage-based with subscription tiers: metered on (1) compute hours of instance runtime and (2) agent credits (1 credit per agent step). Tiers (Free / Pro ~$29 / Business ~$99 / Enterprise custom) bundle compute-hour + credit allotments + concurrency; credit top-ups at $0.04 each, or BYO model-provider API key.

Cost watchouts

Two metered axes (compute hours + agent credits) - long-running or many-step agents consume both; concurrency caps per tier (Free 5 instances); model inference is extra unless you BYO key.

Variable cost rationale

Cost scales with instance runtime (compute hours) x agent steps (credits) x concurrency; BYO provider key shifts model cost off Scrapybara.

Additional watchouts

Usage-based on two axes (compute + credits) - forecast by expected runtime x steps; exact enterprise rates not fully public.

Overage / add-ons

Beyond the tier allotment, you consume compute hours + agent credits; credits top up at $0.04 each (or BYO provider key to offset model cost).

Sales call required

Mixed (some tiers require a call)

Free / trial

n/p

Lowest paid plan

n/p (no public pricing page)

Commercial notes

Computer-use / agent-desktop infrastructure (YC-affiliated); competes with Browserbase, E2B, Anthropic computer-use, Hyperbrowser, Steel.dev

Key ambiguities

Free + Pro ~$29 + Business ~$99 tiers + $0.04/credit + compute-hour model are public; exact enterprise rates + any per-compute-hour overage price not fully captured - verify on the live pricing page.

Cancellation / refund

Free tier to start; self-serve paid tiers (Pro ~$29 / Business ~$99); usage top-ups $0.04/credit; higher-volume/enterprise via sales.

Support SLA / resale

Act SDK (Python) + ComputerTool/BashTool/EditTool; managed orchestration/scaling; works with OpenAI/Anthropic models or BYO key

Missing data

Pricing is public: Free (10 compute hours, 100 agent credits, 5 concurrent instances), Pro ~$29/mo (100 hours, 500 credits, 25 concurrent), Business ~$99/mo (500 hours, 2,500 credits, 100 concurrent), Enterprise custom. Billing meters two axes - compute hours (instance runtime) + agent credits (1 per agent step) - with top-ups at $0.04/credit, or bring your own model-provider API key to offset model cost.

Agentic Index verified 2026-06-25

Alternatives to Scrapybara

The closest documented capability profiles to Scrapybara among browser and computer-use agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.

  • Browser Use11.0 / 14Adds documented Deployment & Data Residency
  • Manus7.5 / 14Adds documented Knowledge Grounding & RAG
  • Stagehand10.5 / 14Adds documented Knowledge Grounding & RAG and Deployment & Data Residency
  • Twin9.5 / 14Adds documented Knowledge Grounding & RAG
  • Airtop10.0 / 14Adds documented Deployment & Data Residency
  • Browserbase12.0 / 14Adds documented Deployment & Data ResidencyScrapybara vs Browserbase →

Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded

Contact us

Found a vendor we missed? Have feedback on the index? We'd love to hear from you.