How much does Capy cost?

No current Scrapybara pricing; the company now sells Capy at capy.ai.

What are the best alternatives to Capy?

Based on documented capability coverage across the Agentic Index 14 point taxonomy, the closest alternatives to Capy among coding agents are Cursor, Google Antigravity, Cline, Cognition. Each is graded on the same public evidence, and no vendor pays for placement.

Does Capy have SOC 2, SSO and role based access control?

Full under the Agentic Index verification standard. Capy publishes public evidence of enterprise security and identity controls, covering a compliance posture such as SOC 2 or ISO 27001, single sign on (SSO or SAML), and role based access control (RBAC). Evidence basis: Capy states SOC 2 Type II (report on trust.capy.ai), SSO with SAML on the Enterprise plan, audit logging, isolated environments, and encryption at every layer. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Can Capy be self hosted or run in a private VPC?

Full under the Agentic Index verification standard. Capy publishes public evidence of self hosted or private VPC deployment, along with data residency options for controlling where agent data is processed and stored. Evidence basis: Agents run in isolated cloud VMs or, through Capy's desktop app, in a folder on the customer's own computer, and the home page offers enterprises on premises hosting or the option to bring their own cloud. Capy does not say where the hosted. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Does Capy give you tracing and observability into agent runs?

Full under the Agentic Index verification standard. Capy publishes public evidence of tracing, observability and audit logging into agent runs, so a team can see what an agent did and why. Evidence basis: Each thread shows the agent's todo plan, terminal output, file changes and diffs, and a status (active, waiting, needs attention, ready for review); tasks show their cost, end to end runs record annotated video, search covers every message. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Back to vendors

Capy

Also known as: Capy AI, Scrapybara, capy.ai

Visit site

Coding agentindependentVerified 2026-09-27

Entry priceNo current Scrapybara pricing; scrapybara.com/pricing no longer existsFull pricing detail

AI software engineer that runs parallel cloud coding agents in isolated VMs, from issue to merged pull request, with Slack, Linear and GitHub integrations. Built by the company behind Scrapybara, whose hosted virtual desktop service shut down in October 2025.

Scrapybara sold hosted Ubuntu, Windows and browser instances that computer use agents could control, with the Act SDK and Python and TypeScript SDKs. Its homepage now says the company built computers for agents and is building Capy, an AI software engineer, at capy.ai, and the Scrapybara pricing page no longer exists. The developer documentation at docs.scrapybara.com still loads.

The capability grid on this record described the hosted instances and is marked unknown until Capy is graded in the lane that fits it. This record's name, address and lane are under review.

Vendor details

Canonical URL

https://capy.ai/

Company status

independent

Use cases & customers

Target customers

developers

Deployment options

SaaSAPI

In practice

You run agents on Scrapybara instances today. Confirm with the vendor whether the service is still sold and supported, since the company now markets Capy.

Sources & related URLs

Research sources

https://scrapybara.com/ https://docs.scrapybara.com/llms.txt https://capy.ai

Agentic Index coverage score

11.5 / 14 capabilities · 82%

Integrations & Tool Calling	Full
Capy acts inside the tools a team uses: it takes Linear issues and streams progress back into them, works in Slack threads, opens and reviews GitHub pull requests, imports Sentry reports, and joins a Tailscale tailnet to reach private databases and internal services. Sourcecapy.airead 2026-09-27
Workflow Orchestration	Full
In Capy a Captain agent plans the work and Build agents carry it out in parallel, each task on its own branch and VM, with a todo list tracking the plan step by step; the product is built for fleets of agents working at once. Scrapybara's hosted desktops shut down in October 2025, and Capy is what the company builds now. Sourcecapy.airead 2026-09-27
Knowledge Grounding & RAG	Partial
Agents work over the customer's repository inside their VM and take in imported context such as Sentry reports and GitHub configs, and attached files land on the agent's machine. Capy describes no maintained retrieval index over the customer's knowledge. Sourcecapy.airead 2026-09-27
Human Oversight & Guardrails	Full
Every task ends in a diff a person reviews before a pull request merges, any agent can be stopped or interrupted mid-run, and a thread moves to needs attention when the agent is blocked on a decision, access or credentials. Nothing shows approval set action by action, though. Sourcedocs.capy.ai/using-capyread 2026-09-27
Security, Identity & Governance	Full
Capy states SOC 2 Type II (report on trust.capy.ai), SSO with SAML on the Enterprise plan, audit logging, isolated environments, and encryption at every layer. Sourcecapy.airead 2026-09-27
Observability & Auditability	Full
Each thread shows the agent's todo plan, terminal output, file changes and diffs, and a status (active, waiting, needs attention, ready for review); tasks show their cost, end to end runs record annotated video, search covers every message both sides wrote, and audit logging is on the Enterprise plan. Sourcedocs.capy.ai/using-capyread 2026-09-27
Memory & State Persistence	Partial
A thread is a durable conversation stored on Capy's servers that holds the full history of its messages, machines, tasks and PRs, survives archiving and resumes when woken, and VM snapshots carry a prebuilt environment. What the agent learns in one thread does not appear to carry into another. Sourcedocs.capy.ai/using-capyread 2026-09-27
Deployment & Data Residency	Full
Agents run in isolated cloud VMs or, through Capy's desktop app, in a folder on the customer's own computer, and the home page offers enterprises on premises hosting or the option to bring their own cloud. Capy does not say where the hosted service runs or whether a region can be chosen. Sourcecapy.airead 2026-09-27
Prebuilt Agents, Templates & Packs	Partial
Three named built in agents ship with Capy: Captain plans, Build executes and Review checks PRs. It also links an automations gallery, but whether that is a catalog of templates a buyer can pick from is unclear. Sourcecapy.airead 2026-09-27
Triggers & Channel Coverage	Full
Automations run on triggers (a Slack message matching a filter, a comment created on a repository) or on cron, and a thread wakes on its own when CI finishes, a review lands or a PR merges, so work starts without a person asking. Sourcedocs.capy.ai/using-capyread 2026-09-27
Model Flexibility & Routing	Full
Capy is model agnostic: a person picks per task from OpenAI, Anthropic, Google, xAI, Moonshot, Zhipu and Alibaba models, can run one task on several models, can bring enterprise API keys, and can run on existing ChatGPT, Copilot or SuperGrok subscriptions. Sourcecapy.airead 2026-09-27
APIs, SDKs & MCP Extensibility	Partial
A REST API with bearer tokens lets a developer start and message threads, interrupt them or archive them. It can also list tasks and launch review agents and VMs. No SDK or MCP server is mentioned. Sourcecapy.airead 2026-09-27
Testing, Debugging & Optimization	Partial
A built-in review agent checks every PR and agents run end to end tests, but those test the customer's code, not the agent. The agent-side tool is running the same task on different models, each with its own history and VM, to compare approaches; no evaluation suite or regression testing of agent behavior is documented. Sourcedocs.capy.ai/using-capyread 2026-09-27
Browser & Computer Use	Full
For end to end testing, Capy's agents use native computer use and browser use inside their own cloud VMs and record annotated videos of each run. Approvals for purchases or form submissions are not described, though a coding sandbox rarely meets either. Sourcecapy.airead 2026-09-27

The Agentic Index coverage score grades every vendor Full, Partial or Not documented 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

Recent platform changes

2026-09-30·Pricing / packagingVerified

Capy 0.4.0 lets teams run the coding agent on AI subscriptions and API keys they already pay for. A new Models page lists a member's and the organization's connections in the order threads use them, a thread moves to the next connection when one runs out, and Capy asks before falling back to its own credits. New connections include several coding plans and AI gateways, and one key from Anthropic, OpenAI, xAI or Google now covers every model from that provider.

Bears on: Pricing / packaging

View source

View all 1 change for Capy →Tracked since Sep 2026 · Verified from public vendor sources

Pricing

No current Scrapybara pricing; scrapybara.com/pricing no longer exists

usage

Included quota

Not published: the Scrapybara pricing page returns 404 and the homepage points to Capy.

What is public

Scrapybara's homepage says the company built computers for agents and is now building Capy, an AI software engineer, at capy.ai. The Scrapybara pricing page no longer exists; the developer docs at docs.scrapybara.com still load.

Billing mechanics

Not published for Scrapybara's hosted instances.

Cost watchouts

Not assessable: no current Scrapybara pricing.

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

The company has moved to Capy; confirm with the vendor whether Scrapybara instances can still be bought before building on them.

Overage / add-ons

Not published.

Sales call required

Mixed (some tiers require a call)

Free / trial

Not published

Commercial notes

Scrapybara, a Y Combinator company, now markets Capy, parallel coding agents, at capy.ai; Capy's organization data links the Scrapybara Y Combinator profile.

Key ambiguities

Whether Scrapybara's hosted instances are still sold or supported after the move to Capy is not stated on the vendor's pages.

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

All Scrapybara pricing. Earlier plan figures (Free, Pro about $29, Business about $99) are withdrawn because no current Scrapybara page publishes them.

Agentic Index verified 2026-09-27

Official pricing page Docs

Related vendors

Alternatives to Capy

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

Cursor12.5 / 14Fuller documented coverage on Prebuilt Agents, Templates & Packs and APIs, SDKs & MCP Extensibility
Google Antigravity12.5 / 14Fuller documented coverage on Prebuilt Agents, Templates & Packs and APIs, SDKs & MCP Extensibility
Cline12.0 / 14Fuller documented coverage on Prebuilt Agents, Templates & Packs and APIs, SDKs & MCP Extensibility
Cognition13.0 / 14Fuller documented coverage on Knowledge Grounding & RAG and Prebuilt Agents, Templates & Packs
OpenAI Codex13.0 / 14Fuller documented coverage on Memory & State Persistence and Prebuilt Agents, Templates & Packs
Poolside11.0 / 14Fuller documented coverage on APIs, SDKs & MCP Extensibility

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

Compare Capy with other vendors →Compare all coding agents →See where Capy ranks →

Data confidence: low