Arcade AI is live · free plan, paid from $25/mo · last verified 22 Oct 2026 · viability healthy
Arcade AI
Arcade AI is the MCP authorization runtime that governs, executes, and audits every action your AI agents take.
Arcade answers the question that stalls agent rollouts: which agent did this, as whom, in which system. Delegated per-user authorization plus one audit record per action is the shortest route through a security review, and a 2,000-event free tier keeps a prototype at zero. Pick it when blast radius is real; pass when your agent only reads public data.
Verdict generatedVerified 8d ago · liveness 79/100 · cite: rightaichoice.com/tools/arcade-ai
- Platform teams shipping agents that act inside real user accounts across enterprise systems
- Security reviewers who need a per-action audit trail naming the agent, the user, and the system
- Enterprises with SOC 2, SSO, and RBAC requirements, including VPC or air-gapped deployments
- Developers who want a governed MCP tool catalog instead of building auth for every integration
- Prototypes and single-user agents calling only public APIs
- Read-only agents where blast radius and per-user accountability aren't a concern
- Teams whose only gap is orchestration and who already have agent authorization solved
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Arcade if your agent runs as a single user against public APIs, or if per-user authorization and audit trails are already solved and all you need is orchestration — LangChain or Composio cover that layer without a runtime.
Every auth event bills at $0.10 on Team, so an agent that re-authenticates users often can outrun the $25/month platform fee in a day.
Fits teams at prototype-to-mid production: Free covers 2,000 auth events and 2,000 tool calls per month, and Team is $25/month plus $0.10 per auth event and $0.01 per tool call, with no per-seat charge. Enterprise is custom, bundling annual auth-event and tool-call volumes with discounted rates. Composio and LangChain are cheaper if you only need orchestration or connection management and will own authorization yourself.
In short
Arcade AI — Arcade AI is the MCP authorization runtime that governs, executes, and audits every action your AI agents take. Best for Platform teams shipping agents that act inside real user accounts across enterprise systems, Security reviewers who need a per-action audit trail naming the agent, the user, and the system, Enterprises with SOC 2, SSO, and RBAC requirements, including VPC or air-gapped deployments. Free to start; paid plans from $25/mo.
What is Arcade AI?
Arcade AI — Arcade AI is the MCP authorization runtime that governs, executes, and audits every action your AI agents take. Best for Platform teams shipping agents that act inside real user accounts across enterprise systems, Security reviewers who need a per-action audit trail naming the agent, the user, and the system, Enterprises with SOC 2, SSO, and RBAC requirements, including VPC or air-gapped deployments. Free to start; paid plans from $25/mo.
Arcade is an actions runtime that sits between your AI agents and every system they reach, enforcing identity, executing tool calls, and logging what happened. It is aimed at platform and security teams whose agents need to act inside real user accounts across Google Workspace, Slack, and Salesforce — and at the reviewers who have to approve that access. The design is identity-first. Authentication runs against your own IdP, and authorization is delegated, so an agent acts as its user and never past that scope. Policies are enforced in the action path itself, outside the model, where a prompt cannot undo them, and credentials stay in the runtime instead of your agent code. On execution you get 8,000+ permission-aware tools plus custom MCP servers you build or bring, and third-party remote MCP servers can be registered under the same control plane. Tool Recommendation exposes the full catalog while cutting token costs up to 37%, built-in evals catch hallucinated and destructive calls before they fire, and Cisco Duo step-up authorization guards sensitive operations. Each action leaves one record naming the agent, the user, and the system, streamed to your SIEM. It is SOC 2 compliant with SSO, RBAC, and audit logs, deployable on Arcade Cloud, in your VPC, on-prem, hybrid, or air-gapped, and it works with any MCP client, model, or framework — Claude Code, Cursor, Codex, VS Code, and Copilot CLI among them. If all you need is orchestration or connection management, LangChain and Composio cover that and leave authorization to you.
Key features
- Agent authorization with delegated per-user permissions
- Authenticate against your own IdP, including Okta, Google, and Microsoft Entra ID
- Policy enforcement in the path of every action, outside the model
- 8,000+ permission-aware tools plus custom MCP servers you build or bring
- Register third-party remote MCP servers and govern their tools from one control plane
- Tool Recommendation exposes the full catalog while cutting token costs up to 37%
- Built-in evals catch hallucinated and destructive calls before they run
- Credentials stay in the runtime and never reach your agent code
- One audit record per action naming the agent, the user, and the system, streamed to your SIEM
- Central control plane to grant, revoke, and version every tool
- Cisco Duo step-up authorization for sensitive actions
- SOC 2 compliant with SSO, RBAC, and full audit logs
- Deploy on Arcade Cloud, in your VPC, on-prem, hybrid, or fully air-gapped
- Install as a plugin in Claude Code, Cursor, Codex, VS Code, and Copilot CLI
- Connect any remote HTTP MCP client via server URL with browser sign-in
Behind the verdict
Where most agent stacks break is not the model call, it is the moment the agent tries to write something. Service accounts flatten everyone to the lowest common denominator, inherited user identity over-provisions, and building your own governance layer means owning the consequence of every action across hundreds of tools. Arcade's bet is that you would not build your own SSO today, so you should not build agent authorization either. We would reach for it when agents touch multi-user enterprise systems and someone in security has to sign off. The delegated-identity model plus a single audit record per action is exactly the artifact a reviewer asks for, and central tool versioning means revoking one integration does not require a redeploy. The 8,000+ tool catalog and Tool Recommendation also cut real token spend — up to 37% per Arcade's numbers — which matters once your catalog gets wide. Pass when the agent only reads public data, runs single-user, or your only gap is orchestration. In those cases the runtime is overhead you will pay for in setup and metering. Teams that already solved authorization in-house won't find a reason to switch either. The closest alternative is Composio, which handles connections and tool orchestration but leaves the authorization model to you; LangChain covers the orchestration half. Arcade's own comparison page names Composio directly, which tells you where they see the fight. Caveats worth modeling before you commit. Usage meters on auth events and tool calls, so a chatty agent at volume gets expensive fast — do the math on monthly tool calls at $0.01 and auth events at $0.10 each before you scale. The runtime also assumes an IdP and a SIEM to plug into; without those, half the value is unreachable. And the per-action audit trail is
Models under the hood
as of 2026-10-22
Limitations
- The Free tier caps you at 2,000 auth events and 2,000 tool calls per month — enough to prototype, not to run production traffic.
- On Team you pay a $25/month platform fee plus metered usage, with auth events at $0.10 each and tool calls at $0.01 each, so cost scales with how chatty your agents are rather than with headcount.
- The governance features security reviews usually hinge on (SSO, RBAC, audit logs, private registry access) and flexible deployment (VPC, air-gapped, hybrid) sit on the Enterprise tier at custom pricing.
- The runtime also assumes you have MCP-connected systems and multiple end users to act for; without those there is little for it to govern.
as of 2026-10-11
Researching Arcade AI? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Is Arcade AI still active?
Arcade AI is live · free plan, paid from $25/mo · last verified 22 Oct 2026 · viability healthy.
Viability score
How well maintained and how widely used is Arcade AI? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →What's new in Arcade AI
Checked 10 days agoAcross the latest 1 update: 1 news mention.
What do users say about Arcade AI?
What people actually say about Arcade AI — is it worth it?
Our community sentiment research written Jul 31, 2026 is more than 90 days old, so it is being re-reviewed against current data and is not shown.
How we verified this
Every listed fact about Arcade AI— pricing, features, links, status — is re-read from the vendor's own pages by our verification pipeline and each change is logged with a date; the last pass ran on 22 Oct 2026. Pages mix model-assisted synthesis with human review; see how we verify.
Verification history
We have re-verified Arcade AI 21 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — changed SSL certificate
- — changed pricing page, who should skip it, pricing tiers, summary, our verdict, features, who it suits, description
- — changed description, pricing tiers, who it suits, summary, our verdict, features, who should skip it
- — changed pricing page
- — changed switching in, summary, who should skip it, our verdict, who it suits, switching out, setup time, hidden costs, description, pricing fit, limitations, pricing tiers
- — changed integrations, features, pricing tiers, our verdict, who it suits, who should skip it, description
Showing the 6 most recent of 21 verification passes.
Free to cite with attribution — this page re-verifies continuously.
How much does Arcade AI cost in 2026?
Arcade AI: Free to start; paid plans from $25/mo. Pricing last verified 22 Oct 2026.
Plans compared
Every published Arcade AI tier, its price as the vendor publishes it, who it fits, and what it adds over the tier below. The 12-month cost projection follows.
| Plan | Price | Billing | Who it's for |
|---|---|---|---|
| Free | Free | — | A developer or small team prototyping a first agent on Arcade Cloud before any production traffic exists.Adds: Free entry point: $0/mo with no credit card, 2,000 auth events and 2,000 tool calls per month, and community support. |
| Team | $25/moVendor: “$25/mo + usage” | monthly | A team that has proven an agent prototype and is moving it into production for real end users.Adds: Adds the $25/month platform fee plus metered usage at $0.10 per auth event and $0.01 per tool call, fully managed on Arcade Cloud with next-business-day email support. |
| Enterprise | Custom | quote | Enterprises where a security review requires SSO, RBAC, audit logs, and a choice of deployment location.Adds: Custom pricing with annual bundles of auth events and tool calls, discounted overage rates, VPC or air-gapped deployment, 24/7 SLA support, and private registry access. |
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Where the pricing makes sense
The company stage and team size where Arcade AI's pricing actually pencils out — and where peers do it cheaper.
Fits teams at prototype-to-mid production: Free covers 2,000 auth events and 2,000 tool calls per month, and Team is $25/month plus $0.10 per auth event and $0.01 per tool call, with no per-seat charge. Enterprise is custom, bundling annual auth-event and tool-call volumes with discounted rates. Composio and LangChain are cheaper if you only need orchestration or connection management and will own authorization yourself.
Has Arcade AI's pricing changed?
No plan price for Arcade AI has moved since we began logging in Aug 2026; the pricing page is re-read on every verification pass.
Who should pick Arcade AI?
Pick Arcade AI if you are Platform teams shipping agents that act inside real user accounts across enterprise systems, Security reviewers who need a per-action audit trail naming the agent, the user, and the system, Enterprises with SOC 2, SSO, and RBAC requirements, including VPC or air-gapped deployments, Developers who want a governed MCP tool catalog instead of building auth for every integration.
Look elsewhere if you are Prototypes and single-user agents calling only public APIs, Read-only agents where blast radius and per-user accountability aren't a concern, Teams whose only gap is orchestration and who already have agent authorization solved, Buyers who won't model metered spend, since tool calls and auth events bill per unit at volume.
Skip it if Skip Arcade if your agent runs as a single user against public APIs, or if per-user authorization and audit trails are already solved and all you need is orchestration — LangChain or Composio cover that layer without a runtime.
GeneratedReal-world workflow fit
Concrete scenarios for the personas Arcade AI actually fits — and what changes day-one when you adopt it.
You wire an internal support agent to Slack and Salesforce. Auth runs against Okta, so each rep's agent acts inside that rep's own scope; built-in evals block a destructive delete before it fires.
Outcome: The agent ships to the support team without a custom permission model, and every ticket update leaves a SIEM-streamed record naming the agent, the user, and the system.
You register a partner's remote MCP server under Arcade's control plane, put Cisco Duo step-up authorization on the sensitive tools, and point the audit stream at your existing SIEM.
Outcome: Third-party agent tools go through the same governance path as your own, so security reviews one runtime instead of every integration.
You start on the Free tier with 2,000 auth events and 2,000 tool calls to prove the flow, then move to Team at $25/month plus $0.10 per auth event and $0.01 per tool call when reps actually use it.
Outcome: The agent takes secure actions on a rep's behalf in Gmail and Salesforce without you storing refresh tokens or building auth plumbing.
Use cases
- Automate CRM updates by letting an AI agent read and write Salesforce records on behalf of sales reps.
- Enable an AI assistant to send emails via Gmail with user-specific authorization without storing tokens.
- Build a support agent that queries Slack history and creates tickets with per-user permissions.
- Deploy an onboarding agent that reads Google Calendar and sends personalized welcome messages via Slack.
- Create a data analysis agent that accesses Google Sheets and generates reports with user-specific view permissions.
- Power an AI sales agent that analyzes calls, emails, and CRM data and takes secure actions on behalf of a rep.
- Connect an IDE agent such as Cursor to external APIs through Arcade MCP tools without storing API keys in the IDE.
- Bring a partner or vendor's remote MCP server under your own control plane with per-user authentication.
Setup time & first value
How long it actually takes to get something useful out of Arcade AI — broken out by persona, not the marketing-page minute.
A developer can connect an IdP and run a first governed action on the Free tier in an afternoon — Arcade Cloud is fully managed and no credit card is required. Production rollout is longer: expect days to weeks to model policy, choose deployment (Cloud, VPC, on-prem, hybrid, or air-gapped) and get security sign-off, with Enterprise environments also waiting on SSO, RBAC, and private registry
Switching to or from Arcade AI
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From hand-rolled auth and token storage: point authentication at your IdP and let Arcade hold credentials in the runtime instead of your app.
- →From Composio or LangChain tool orchestration: keep the orchestration layer and move authorization into Arcade's control plane in front of it.
- →From service-account agents: swap shared credentials for delegated per-user authorization so each action ties back to a real user.
- →From direct MCP server connections: register the servers in Arcade and govern their tools from one control plane with per-user auth.
- ↗To LangChain or Composio: keep the tools and move only orchestration out, then rebuild the authorization and audit layer you were relying on Arcade for.
- ↗To a self-built governance layer: export the tool registry and re-implement policy enforcement in the path of each action yourself.
- ↗To direct provider APIs: replace MCP tool calls with per-provider SDK calls and handle token refresh and per-user scoping in your own code.
Integrations
Tutorials & Learning
YouTube returned 6 videos for “Arcade AI”, and we withheld 6: 6 could not be judged, because “Arcade AI” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about Arcade AI.
Official links
What are the best Arcade AI alternatives?
The closest alternatives to Arcade AI we list are Kastra, Kontext Cli and Composio. Ranked by direct product-type match, not category overlap.
| Alternative | Pricing | Starts at | Last verified |
|---|---|---|---|
| Kastra | Freemium | Pricing not published | — |
| Kontext Cli | Freemium | Pricing not published | — |
| Composio | Freemium | Pricing not published | — |
Alternatives to Arcade AI
View all
Kastra
Runtime authorization for AI agents — allow, deny, redact, or hold every action before it executes
Kontext Cli
Kontext is runtime authorization for AI agents — it hooks every Claude Code, Codex, and Cowork tool call before it runs, with no gateway in
Tools that pair well with Arcade AI
Common stack mates teams adopt alongside Arcade AI, with the specific reason each pairing earns its keep.
Kastra
Runtime authorization for AI agents — allow, deny, redact, or hold every action before it executes
Kontext Cli
Kontext is runtime authorization for AI agents — it hooks every Claude Code, Codex, and Cowork tool call before it runs, with no gateway in
Composio
Composio gives AI agents managed tool-calling across 1,600+ apps with per-user auth and secure execution.
Arcade AI vs its closest rival
Arcade AI's closest head-to-head is with Statewave — Pick Statewave if your bottleneck is agents forgetting context and you want to self-host a memory layer you fully control. Pick Arcade AI if your bottleneck is securely connecting agents to real user accounts and enterprise tools — it ships auth, governance, and a huge MCP catalog out of the box. Most teams shipping production agents today will find Arcade's faster time-to-value worth the trade-off, unless you have a very specific memory-heavy use case. Verdict generated
Head-to-head comparisons
Statewave vs Arcade Ai
Pick Statewave if your bottleneck is agents forgetting context and you want to self-host a memory layer you fully control. Pick Arcade AI if your bottleneck is securely connecting agents to real user accounts and enterprise tools — it ships auth, governance, and a huge MCP catalog out of the box. Most teams shipping production agents today will find Arcade's faster time-to-value worth the trade-off, unless you have a very specific memory-heavy use case.
GeneratedNpm I G Hotcell vs Arcade Ai
If you need airtight local control and privacy for agent experiments, hotcell is your choice. But if you're shipping agents that act in real user accounts across enterprise tools like Salesforce or Slack, Arcade AI's pre-built auth, governance, and MCP tools will save you months — and its SOC 2 compliance makes the security review a non-event. Pick hotcell for sandboxed iteration, Arcade for production.
GeneratedVoicemem vs Arcade Ai
These aren't competitors — you would not swap one for the other, and a comparison page is the wrong place to decide. Arcade AI is a buy-an-enterprise-runtime decision for teams whose agents touch real user accounts and need per-action audit trails naming agent, user, and system. VoiceMem is a self-host, Apache-2.0 memory layer for people building a real-time voice agent who want persona and emotion recall at low latency, and who accept v0.0.2 maturity with no support. If you're asking "which of these do I pay for," the answer depends entirely on whether your problem is authorization or memory — and if it's both, you'd run them in different parts of your stack, not pick between them.
GeneratedFrequently asked questions
Best-of guides
Topics
Used Arcade AI? Help shape our editorial sentiment research.