Simular
AI agent that operates desktop and web software autonomously to complete knowledge work tasks.
Simular is a computer-use AI company building agents that operate a computer the way a person does, by looking at the screen and controlling the mouse and keyboard, rather than by calling APIs. Founded by a team of researchers who have worked on AI agents since well before the current wave, Simular's premise is that the large amount of work living inside graphical interfaces and desktop software, not clean programmatic interfaces, can be handed to an agent that simply drives the computer. The company frames itself as building toward an autonomous computer that does the work so people do not have to.
Underneath is Agent S, Simular's open-source agentic framework that uses computers like a human through an agent-computer interface. It is a research-driven project, refined across successive generations, that plans hierarchically and draws on past experience to carry out multi-step tasks across desktops, browsers, and mobile, and it has been notably strong on the standard computer-use benchmarks that span Windows, Linux, and Android environments. The framework is open and available for developers to build on.
On top of that research sits Sai, Simular's commercial agent and the part most users interact with. Sai operates a Mac or Windows desktop, a browser, and an Android phone, clicking, typing, and navigating apps, and it can also use terminals, call APIs, and write code when that is the better path. It runs either locally on a machine you already own, for privacy, or inside a private, isolated cloud virtual desktop that stays always on, so it can keep working unattended without touching your real environment.
Because driving the full native desktop is more powerful and more consequential than operating a single hosted web page, Simular leans hard on safety: Sai is built with guardrails and pauses to confirm with the user before taking sensitive or critical actions rather than running fully hands-off. This makes it suited to operators and knowledge workers who want to delegate repetitive desktop chores, moving data between applications, filling forms, and running routine multi-step processes, to an agent that works across the same apps and devices they already use.
Vendor details
Canonical URL
https://www.simular.ai
Category
Browser / computer-use agent
Company status
independent
Use cases & customers
Target customers
Deployment options
In practice
Your repetitive work lives in desktop apps and terminals, not a single web page, so browser-only agents can't help. Simular's Sai drives the full Mac or Windows desktop, clicking and typing like a person to finish the task.
You want an agent that can keep working while you're away but not touch your real machine. Sai runs in a private, isolated cloud virtual desktop, always on, so unattended jobs don't risk your own environment.
You're building computer-use agents and want a proven open framework to start from. Simular's open-source Agent S plans hierarchically and operates GUIs across desktop, browser, and mobile, a research base you can build on.
Capability coverage
7.5 / 14 capabilities · 54%
| Integrations & Tool CallingAgent Features research report + JSON Feature Rubric | Partial |
|---|---|
| Workflow OrchestrationAgent Features research report + JSON Feature Rubric | Full |
| Knowledge Grounding & RAGAgent Features research report + JSON Feature Rubric | Unable to verify |
| Human Oversight & GuardrailsAgent Features research report + JSON Feature Rubric | Full |
| Security, Identity & GovernanceAgent Features research report + JSON Feature Rubric | Partial |
| Observability & AuditabilityAgent Features research report + JSON Feature Rubric | Full |
| Memory & State PersistenceAgent Features research report + JSON Feature Rubric | Partial |
| Deployment & Data ResidencyAgent Features research report + JSON Feature Rubric | Full |
| Prebuilt Agents, Templates & PacksAgent Features research report + JSON Feature Rubric | Unable to verify |
| Triggers & Channel CoverageAgent Features research report + JSON Feature Rubric | Unable to verify |
| Model Flexibility & RoutingAgent Features research report + JSON Feature Rubric | Unable to verify |
| APIs, SDKs & MCP ExtensibilityAgent Features research report + JSON Feature Rubric | Partial |
| Testing, Debugging & OptimizationAgent Features research report + JSON Feature Rubric | Partial |
| Browser & Computer UseAgent Features research report + JSON Feature Rubric | Full |
Recent platform changes
Simular introduced Simular Pro as its first production-ready computer-use agent.
View sourcePricing
From $20/mo per computer · 7-day trial
hybrid
Included quota
Plus ($20/mo/computer, 7-day free trial / limited access): reloadable credits (~10,000 per pricing page), 1 developer agent, macOS/Windows, Workflow Editor (Developer Mode), Community Gallery workflow sharing, community support - for personal projects/scripting. Pro ($500/mo/computer): everything in Plus + unlimited credits, 1 deployment agent, virtual-machine support, private/team workflow sharing, full Simulang support, zero LLM-provider data retention, priority engineering support - for business ops/back-office/IT. Enterprise (custom): everything in Pro + managed scaling, all platforms/VMs/private cloud, advanced editor + custom extensions, audit-log workflow sharing, SOC2/HIPAA/SSO/RBAC, dedicated account + engineering team.
What is public
Simular (simular.ai - 'the autonomous computer company'; its product Sai is a 24/7 AI coworker that operates Windows/macOS/Linux/VMs like a human, WITHOUT APIs, via clicks/keystrokes) prices PER COMPUTER, monthly. Current tiers (per the live pricing page): Plus $20/mo per computer ($200 struck through as an early-access discount; reloadable credits - the pricing page cites 10,000 credits while the sai.work microsite frames it as '$20 credits/mo'; 1 agent; now offered with a 7-day free trial / 'Limited Supply' rather than strictly invite-only), Pro $500/mo per computer (unlimited credits, VM support, 1 deployment agent, zero LLM data retention), and Enterprise (contact for price; SOC2/HIPAA/SSO/RBAC, managed scaling). No annual pricing yet. The underlying Agent S agent framework is open-source/free on GitHub.
Billing mechanics
Billed per COMPUTER (device) per month, with a CREDIT layer on the entry tier: Plus ($20/mo/computer) includes reloadable credits (10,000 per the pricing page, or '$20 credits/mo' per the sai.work microsite - framing varies); Pro ($500/mo/computer) and Enterprise carry UNLIMITED credits (so the per-computer fee is effectively flat at those tiers). Agents run on your own device or on cloud virtual machines (Pro+). You can switch/upgrade/cancel anytime; no annual billing currently. A 7-day free trial is now offered on Plus. A referral program pays $10/signup (Plus) and $100/pro-signup (Pro).
Cost watchouts
Pricing is PER COMPUTER, so multi-device automation multiplies the subscription; the jump from Plus ($20) to Pro ($500/mo/computer) is steep (25x) - driven by unlimited credits + VM support; Plus access is limited/invite-gated (a 7-day free trial is now offered; an invite code or reposting their launch video can get you ahead); no annual pricing/discount yet; credit-consumption rates per task aren't published; the commercial product has no permanent free tier (the open-source Agent S is the free path)
Variable cost rationale
Per-computer flat at Pro/Enterprise (unlimited credits = predictable) but credit-bounded on Plus (reloadable credits, so heavy use means reloads); the real variable is device count x tier, not per-task usage at the upper tiers
Additional watchouts
Per-computer pricing scales with device count; the Plus-to-Pro gap is huge ($20 to $500/mo/computer); Plus is invite-gated (7-day trial now offered); no annual discount yet; per-task credit consumption is unpublished (hard to forecast Plus's credit runway); the seed's older device/server-hours pricing is superseded - verify current tiers on the live page
Overage / add-ons
Plus is credit-bounded (reloadable credits - you reload when depleted); Pro and Enterprise have unlimited credits, so the per-computer fee is the cap. Scaling to more devices means more per-computer subscriptions; cross-platform/VM/private-cloud scaling moves you to Pro or Enterprise.
Sales call required
Mixed (some tiers require a call)
Free / trial
Free; Sai 7-day trial
Lowest paid plan
$19.99/device/mo
Commercial notes
Research-driven (team researching AI agents since 2017); claims to top benchmarks across browsers, computers, AND smartphones simultaneously; open-source Agent S / S2 / S3 lineage (github.com/simular-ai/agent-s); the key differentiator is API-free desktop automation (operates the GUI like a human across Win/Mac/Linux/VMs); product 'Sai' positioned as an always-on cloud-virtual-desktop coworker; competes with Manus and the wave of computer-use agents
Key ambiguities
Simular's pricing was RESTRUCTURED - the current live page shows Plus $20/mo/computer (reloadable credits) / Pro $500/mo/computer (unlimited) / Enterprise custom, which SUPERSEDES the older 'Premium $19.99/device + Server add-on $39.99/device (200 hrs, then $0.10/hr)' structure; the Plus headline shows '$200' struck through to '$20' (an early-access/launch discount); Plus now includes a 7-day free trial; the Plus credit allotment is cited inconsistently (10,000 credits on the pricing page vs '$20 credits/mo' on sai.work); no permanent free commercial tier (vs the older 'Free' tier); per-task credit costs aren't published
Cancellation / refund
Switch/upgrade/downgrade/cancel anytime from the dashboard; Plus now offers a 7-day free trial; no annual pricing currently (reach out for long-term partnership); Plus access is limited/invite-gated (invite code, or get ahead by reposting the launch video); Enterprise is custom-contracted; the open-source Agent S framework is free (GitHub) with no contract
Support SLA / resale
Plus: community support (Discord); Pro: priority email/chat/Slack support + zero LLM-provider data retention; Enterprise: dedicated onboarding + engineering team, SOC2/HIPAA/SSO/RBAC, audit logs, private cloud; agents run on-device or on cloud VMs (Pro+); workflow sharing via public Community Gallery (Plus) or private/team (Pro/Enterprise); contact support@simular.ai
Missing data
Per-task credit-consumption rates aren't published, so the Plus credit runway is hard to estimate; no annual pricing yet. Current plans: Plus $20/mo per computer ($200 list, struck to $20 early-access; reloadable credits, 7-day free trial, limited/invite access), Pro $500/mo per computer (unlimited credits + VM support), Enterprise custom (SOC2/HIPAA/SSO/RBAC). The open-source Agent S framework remains free on GitHub. The live pricing page is actively changing - verify before relying on figures.
Related vendors
- AgentQL — A self-healing query language for the web: natural language…
- Airtop — Cloud browsers for AI agents that compile plain English workflows…
- Asteroid — Supervised browser and computer use agents for regulated back…
- Astral — Browser based AI marketing agent that operates social platforms like…
- Automat AI — Managed AI agents that operate computers visually to replace legacy…
- Autotab — General AI agent that uses a mouse and keyboard like a person to do…