How much does Simular cost?

Sai Starter $50 a month (list $200), Premium $200, Pro $500, Enterprise by quote; 7 day free trial; Agent S and Simulang are open source.

What are the best alternatives to Simular?

Based on documented capability coverage across the Agentic Index 14 point taxonomy, the closest alternatives to Simular among browser and computer-use agents are Skyvern, Browserbase, Automat AI, Browser Use. Each is graded on the same public evidence, and no vendor pays for placement.

Does Simular support MCP and tool calling?

Full under the Agentic Index verification standard. Simular publishes public evidence of Model Context Protocol (MCP) support, tool calling, and API or SDK extensibility for connecting agents to external tools. Evidence basis: Simular's own SDK, simulang-js, has a versioned class and function reference (Machine, Instance, AccessibilityTree, Screen, GroundingModel and more), a simulang CLI and a Simulang skill for Claude Code. Agent S installs as the gui-agents Py. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

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

Full under the Agentic Index verification standard. Simular 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: The Enterprise plan includes SSO, RBAC and SOC 2. The attestation's type, date and report are not published, and there is no trust center. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

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

Full under the Agentic Index verification standard. Simular 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: Sai runs either on a private virtual machine Simular provisions or on the customer's own Mac, Windows machine or Mac mini, and Agent S and Simulang run on the customer's machines. No region list is published, and Simular does not say where. Graded from public evidence only against the Agentic Index 14 point capability taxonomy. No vendor pays for placement.

Back to vendors

Simular

Visit site

Browser / computer-use agentindependentVerified 2026-09-27

Entry priceFrom $50 a month (Starter) · 7 day free trialFull pricing detail

Computer use agent company: Sai operates desktop apps and websites on a cloud or local computer, beside the open source Agent S and the Simulang scripting SDK.

Simular builds agents that operate a computer the way a person does, through the screen, keyboard and mouse, not through APIs. It sells Sai, its commercial computer use agent, and publishes two open source projects, Agent S, a framework for agents that control macOS, Linux and Windows, and Simulang, a scripting layer for desktop and browser automation.

Sai takes a plain language request and works through apps, websites and desktop interfaces, and it can also call APIs, use the terminal and write code, with skills to extend it. It runs on a private virtual machine Simular provisions, which is always on at the Premium tier, or on the customer's own Mac, Windows machine or Mac mini. Sai checks with the user before critical actions and shows its work live. A Zoom integration lets it schedule and manage meetings, and a workflow templates library offers ready to run tasks, such as invoice processing into Google Sheets, resume screening and a daily email summary, each with its prompt and inputs.

For developers, Simulang is a TypeScript and JavaScript SDK with a CLI and a Claude Code skill. Scripts drive real apps through the accessibility tree, fall back to vision grounding, call models for decisions and compose into larger workflows. Agent S installs as a Python package and runs on the model the developer chooses, including OpenAI, Anthropic, Azure OpenAI, Groq, Ollama and vLLM.

Sai is sold on monthly plans from Starter to Pro, with SSO, RBAC and SOC 2 listed on the quoted Enterprise plan. There is no run history or audit log, no scheduled or event trigger, and no memory layer.

Vendor details

Canonical URL

https://www.simular.ai

Company status

independent

Use cases & customers

Target customers

knowledge workersenterprises

Deployment options

desktopSaaS

Integrations

Zoom through Sai's Settings, Integrations page (schedule and manage meetings, recordings); Sai also operates any app or website through its interface. Simulang has a Claude Code skill; Agent S works with OpenAI, Anthropic, Azure OpenAI, Groq, Ollama and vLLM.

In practice

You have repetitive desktop work in apps with no API, such as moving invoice data into a spreadsheet. Sai does it on a cloud computer or your own Mac or PC, and checks with you before critical actions.

You want a computer use agent you can run and modify yourself. Agent S is open source, runs on macOS, Linux or Windows and uses the model provider you choose.

Your team scripts desktop automation. Simulang lets you write TypeScript that drives real apps through the accessibility tree, with vision grounding when the tree is not enough, and run it from Claude Code.

Sources & related URLs

Research sources

https://www.simular.ai https://www.simular.ai/sai https://www.simular.ai/pricing https://www.simular.ai/workflow-templates https://docs.simular.ai/llms.txt https://docs.simular.ai/introduction https://docs.simular.ai/simulang/simulang-primer https://docs.simular.ai/agent-s/quickstart https://docs.simular.ai/sai/integrations/zoom

Agentic Index coverage score

10.0 / 14 capabilities · 71%

Integrations & Tool Calling	Partial
Sai connects to Zoom by OAuth from its Settings, Integrations page, so it can schedule and manage meetings and fetch recordings. Zoom is the only named connector. Sai also calls APIs, runs the terminal and supports skills, but there is no custom tool mechanism. Sourcedocs.simular.ai/sai/integrations/zoomread 2026-09-27
Workflow Orchestration	Full
Simulang scripts in TypeScript or JavaScript mix deterministic interface steps with model steps (vision grounding, AskModel, DecisionModel, StateSatisfiesModel). They wait and retry on changing pages and compose larger workflows from imported helper modules and scripts run as subprocesses. Sourcedocs.simular.ai/simulang/simulang-primerread 2026-09-27
Knowledge Grounding & RAG	Not documented
There is no retrieval over the customer's documents and no maintained knowledge index. Agent S can optionally retrieve web knowledge through a self hosted Perplexica search, which is web search at run time, not the customer's knowledge. Sourcedocs.simular.ai/agent-s/quickstartread 2026-09-27
Human Oversight & Guardrails	Full
Sai uses approval based control for important actions and double checks with the user before it executes critical actions. Simular does not say which actions this covers. Sourcesimular.ai/sairead 2026-09-27
Security, Identity & Governance	Full
The Enterprise plan includes SSO, RBAC and SOC 2. The attestation's type, date and report are not published, and there is no trust center. Sourcesimular.ai/pricingread 2026-09-27
Observability & Auditability	Partial
Sai shows the agent working live in its dedicated workspace. There is no run history, step trace, audit log or export, and Simulang's logger is the script author's own logging. Sourcesimular.ai/sairead 2026-09-27
Memory & State Persistence	Partial
The Premium plan gives Sai an always on Windows or Mac computer that keeps its files and state between tasks. That is the machine's state, not a memory layer, and there is no memory store, scope or lifetime. Sourcesimular.ai/pricingread 2026-09-27
Deployment & Data Residency	Full
Sai runs either on a private virtual machine Simular provisions or on the customer's own Mac, Windows machine or Mac mini, and Agent S and Simulang run on the customer's machines. No region list is published, and Simular does not say where model traffic is processed. Sourcesimular.ai/sairead 2026-09-27
Prebuilt Agents, Templates & Packs	Full
A workflow templates library offers ready to run tasks for Sai, each with the exact prompt, the inputs it needs and what it returns. They cover separate jobs such as invoice processing with OCR into Google Sheets, resume screening and a daily email summary. The templates are prompts the customer edits and runs, not packaged agents. Sourcesimular.ai/workflow-templatesread 2026-09-27
Triggers & Channel Coverage	Partial
Work reaches Sai as a plain language request in the Sai web or desktop app, and Simulang scripts run when a developer or Claude Code starts them. There is no schedule, webhook or event trigger. Sourcesimular.ai/sairead 2026-09-27
Model Flexibility & Routing	Full
The open source Agent S runs on the model the customer sets, from OpenAI, Azure OpenAI, Anthropic, Groq, Ollama or vLLM, with the customer's own keys. The commercial Sai offers no model choice. Sourcedocs.simular.ai/agent-s/quickstartread 2026-09-27
APIs, SDKs & MCP Extensibility	Full
Simular's own SDK, simulang-js, has a versioned class and function reference (Machine, Instance, AccessibilityTree, Screen, GroundingModel and more), a simulang CLI and a Simulang skill for Claude Code. Agent S installs as the gui-agents Python package, and the Pro plan includes full API access to Sai. The openapi.json in the docs is a placeholder sample. Sourcedocs.simular.ai/introductionread 2026-09-27
Testing, Debugging & Optimization	Not documented
The OSWorld results on the homepage are Simular's own benchmark, and Simulang's troubleshooting guidance (printing the tree, iterating in the REPL) is debugging. There is no evaluation harness, scored test cases or quality gate for the customer's agents or workflows. Sourcedocs.simular.ai/llms.txtread 2026-09-27
Browser & Computer Use	Full
Sai operates apps, websites and desktop interfaces on a private virtual machine Simular provisions or on the customer's own Mac or Windows machine. The open source Agent S controls macOS, Linux and Windows through the interface, keyboard and mouse, and Simulang scripts drive real apps through the accessibility tree, with vision grounding as a fallback. Sourcesimular.ai/sairead 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

Product updates

September 30, 2026·MCP / tool calling / APIVerified

A new MCP server from Simular lets Claude Code, Codex, Cursor or any MCP client hand plain English tasks to Sai, its computer use agent, which carries them out on a Windows cloud computer Simular hosts. Its tools cover starting, approving, waiting on and stopping tasks, plus file uploads.

Bears on: Browser/computer use

View source

July 8, 2025·Browser/computer useVerified

Simular introduced Simular Pro as its first production-ready computer-use agent.

Bears on: Browser / computer-use agent

View source

View all 2 Simular updates →Tracked since Jul 2025 · Verified from public vendor sources

Pricing

From $50 a month (Starter) · 7 day free trial

Monthly plans, with a credit allowance on Starter and Premium, more credits for purchase at any time, and unlimited use within fair use limits on Pro. Enterprise is quoted.

Trial available

Included quota

Starter, at $50 a month against a $200 list price, includes expanded access to a cloud Windows computer, credits with more for purchase, bring your own Windows or Mac device, and community support. Premium, at $200 a month, adds an always on Windows or Mac computer, a higher monthly credit allowance and priority support. Pro, at $500 a month, adds unlimited usage within fair use, team workflow sharing, full API access, and highest priority support and capacity. Enterprise is quoted and adds managed scaling, a dedicated account team, SSO, RBAC and SOC 2, custom integrations and SLAs.

What is public

Simular publishes four Sai plans at simular.ai/pricing, with Starter at $50 a month (shown against a $200 list price), Premium at $200 a month, Pro at $500 a month and Enterprise by quote. Sai comes with a 7 day free trial. Agent S and Simulang are open source on GitHub.

Billing mechanics

Plans are billed monthly, with a credit allowance on Starter and Premium and more credits for purchase at any time. Pro is unlimited within fair use limits, and Enterprise is quoted. Whether plans are priced per computer is not stated.

Cost watchouts

Credit top ups add cost on Starter and Premium. Unlimited use and API access require the jump from Premium ($200) to Pro ($500). Security controls sit only on Enterprise.

Variable cost rationale

Starter and Premium meter credits. Each plan carries a monthly allowance, and heavy use means buying more credits at any time, at a credit cost per task that is not published. Pro removes the meter, with unlimited usage within fair use limits for $500 a month, so moving up to Pro caps a heavy user's bill. Enterprise is quoted.

Additional watchouts

Starter's $50 is shown against a $200 list price, and how long the discount lasts is not stated. Credit consumption per task is not published, so Starter and Premium runway is hard to forecast. SSO, RBAC and SOC 2 sit on the quoted Enterprise plan only.

Overage / add-ons

Starter includes credits, and more can be bought at any time. Premium carries a higher monthly allowance. Pro is unlimited within fair use limits, and Enterprise is quoted.

Sales call required

Mixed (some tiers require a call)

Free / trial

7 day free trial of Sai. Agent S and Simulang are free and open source

Lowest paid plan

Starter at $50 a month, shown against a $200 list price

Commercial notes

Simular makes Sai, a computer use agent sold on monthly plans, and publishes Agent S, an open source computer use framework, and Simulang, a scripting SDK for desktop automation. A workflow templates library offers ready to run Sai tasks.

Key ambiguities

Starter is $50 against a struck $200, with no word on how long that price holds. Credit amounts and whether plans are per computer are not stated.

Cancellation / refund

Plans can be upgraded, downgraded or cancelled anytime from the account dashboard. Enterprise is contracted by quote.

Support SLA / resale

Starter has community support, Premium priority support, and Pro highest priority support and capacity. Enterprise adds a dedicated account team and SLA guarantees.

Missing data

Credit amounts per plan and the credit cost of a task are not published. Annual pricing is not shown, and whether a plan covers one computer or several is not stated.

Agentic Index verified 2026-09-27

Official pricing page Docs

Related vendors

Alternatives to Simular

The closest documented capability profiles to Simular 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.

Skyvern11.5 / 14Adds documented Testing, Debugging & Optimization
Browserbase10.0 / 14Adds documented Testing, Debugging & Optimization
Automat AI11.5 / 14Adds documented Knowledge Grounding & RAG and Testing, Debugging & Optimization
Browser Use11.5 / 14Adds documented Testing, Debugging & Optimization
Notte9.5 / 14Adds documented Testing, Debugging & Optimization
OpenClaw13.5 / 14Adds documented Knowledge Grounding & RAG and Testing, Debugging & Optimization

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 Simular with other vendors →Compare all browser and computer-use agents →See where Simular ranks →

Data confidence: high

Head to head

Simular vs AutotabSimular vs H Platform