Comparison
actionlake vs Arcade
Arcade alternative for AI agents
Arcade is an MCP runtime that lets agents act as the signed-in user across 8,000+ permission-aware tools, served through MCP gateways and billed at $0.01 per tool call plus $0.10 per auth event on its Team plan. actionlake is an MCP gateway with 995 vendor connectors, billed per upstream call, with spending caps per agent key, human approval before writes and one-call answers across two systems.
Checked 22 October 2026 · every fact links its source
Which to choose
It depends on the job.
Arcade sits between agents and the systems they reach. An MCP gateway federates tools from several MCP servers behind one URL, users sign in through Arcade or your identity provider, and each action runs within both the user's and the agent's permissions. Custom tools are built with Arcade's open-source framework in Python or JavaScript.
Choose Arcade if
- Each action must run as the individual user, through your identity provider, within that user's own permissions.
- You build custom MCP tools and want an open-source framework and a private registry.
- You need VPC or air-gapped deployment, or SOC 2 from the vendor today.
- You want tool recommendation to keep a large tool set out of the agent's context.
Choose actionlake if
- You want a spending cap per agent key and a person's approval before a write runs.
- You ask one question across two systems and want the joined answer in one call.
- You run high volume: Arcade's published Team rate is $0.01 per tool call.
- Your agent connects to business systems with API keys it can set up itself, without a sign-in per user.
Credit where due
Where Arcade is stronger
- Delegated user authorization: the agent acts as its user, never past its own scope, with auth against your identity provider.
- Breadth: 8,000+ tools, plus custom MCP servers you build or bring.
- Deployment: Arcade Cloud, your VPC, or fully air-gapped on Enterprise.
- Certification: Arcade states SOC 2; actionlake holds no certification yet.
The difference
Where actionlake differs
- Arcade bills tool calls and auth events; actionlake bills upstream calls, and searching, checking and cached answers are free.
- actionlake caps spend per key and can hold a write for approval. Arcade describes policy hooks that can block or redact, not a person's approval.
- actionlake connections belong to the tenant; Arcade is built around each end user's own sign-in.
Feature by feature
15 things we compared.
Each Arcade cell links the page we read it on. Checked 22 October 2026.
| Feature | actionlake | Arcade |
|---|---|---|
| Hosted MCP server | YesOne endpoint, https://connect.actionlake.com/mcp, over Streamable HTTP; a REST API with the same behavior. | YesGateways at https://api.arcade.dev/mcp/{slug}, for clients that support Streamable HTTP.Source: docs.arcade.dev/en/guides/mcp-gateways |
| How a client signs in | YesOAuth 2.1 with dynamic client registration and PKCE, or an actionlake key. The agent holds an actionlake key, never a vendor token. | YesArcade Auth, a User Source (OAuth or OIDC with your identity provider), or Arcade headers with an API key and user ID.Source: docs.arcade.dev/en/guides/mcp-gateways |
| Apps and actions | Yes995 vendor connectors and 289,000+ actions behind eight tools. An agent can draft, test and sign a connector for an API that is missing. | Yes8,000+ permission-aware tools; custom MCP servers you build or bring.Source: arcade.dev |
| What is billed | YesOne upstream call: a request actionlake sent to a vendor that returned a definitive response. Searching, checking, planning, cached answers and failures you did not cause (a vendor's 401, 403, 429, 5xx or a timeout) are free. | YesTool calls and auth events: $0.01 per tool call and $0.10 per auth event on Team.Source: arcade.dev/pricing |
| Free allowance | Yes100,000 upstream calls a month, no card. | Yes2,000 tool calls and 2,000 auth events a month, no card.Source: arcade.dev/pricing |
| Limit which actions an agent can use | YesPer key: by action, by service, by verb, or read-only. A key can mint narrower keys, never wider ones. | YesChoose the tools per gateway; each action runs within both the user's and the agent's scope.Source: docs.arcade.dev/en/guides/mcp-gateways |
| Spending cap per agent key | YesA cap in upstream calls per key. A child key's calls count against each parent. Refused before the request is sent. | Not statedNot described on the pages read.Source: arcade.dev/pricing |
| A person approves a write before it runs | YesDestructive actions need an explicit confirmation, and under the key's `human` policy the agent cannot give it, so a person approves. Approval rules make chosen actions, or every write, wait for a named person, who decides in the console or from an emailed signed link. | Not statedPre- and post-tool-call hooks can block sensitive actions; a person's approval step was not described.Source: arcade.dev |
| Record of each call | YesOne ledger row per run, hash-chained and sealed daily, with the agent's stated reason (kept 90 days). Request and response traces are kept for the plan's window. | YesOne record per action naming agent, user and system, streamed to your SIEM. The pricing page lists audit logs under Enterprise.Source: arcade.dev/pricing |
| Scan call data for sensitive content | NoSecrets are removed from traces. Call data is not scanned or redacted for personal data. | PartlyHooks can redact personal data; written by you as policy.Source: arcade.dev |
| One question across two systems | YesThe ask tool fetches from two or more systems, joins on a shared entity such as the customer, and returns the answer, not the rows. | Not statedNot described on the pages read.Source: arcade.dev |
| Event triggers and webhooks | NoBy design. Agents and schedules start the work; a scheduled plan can ask what changed since the last run. | Not statedNot described on the pages read.Source: arcade.dev |
| Self-host | NoHosted service only: actionlake runs in one region in Sydney, Australia, and no self-hosted edition is offered. | YesEnterprise: your VPC or fully air-gapped.Source: arcade.dev/pricing |
| Security certifications | NoNone held yet. SOC 2 Type II, ISO 27001 and a third-party penetration test are in progress. | YesSOC 2, as stated by Arcade.Source: arcade.dev |
| Where data is processed | YesOne region: Sydney, Australia (DigitalOcean syd1). Stored in Australia at rest; TLS terminates at Cloudflare's edge. | Not statedNot stated on the pages read.Source: arcade.dev |
"Not stated" means the vendor pages we read did not describe the feature. It is not a claim that the feature is missing. Checked 22 October 2026.
Pricing
What 100,000 and 1,000,000 calls a month cost.
Arcade's unit: Tool call, plus auth events billed separately. Plan used: Team, $25 a month plus usage. actionlake's figure is the cheapest plan that fits the volume, from the published price book; one upstream call counts as one action.
| Calls a month | actionlake | Arcade |
|---|---|---|
| 100,000 | US$0 a monthUS$0.00 per 1,000 callsFree: US$0 a month, 100,000 calls included. | US$1,025 a monthUS$10.25 per 1,000 calls$25 + 100,000 tool calls at $0.01. Auth events not included.Source: arcade.dev/pricing, read 22 October 2026 |
| 1,000,000 | US$219 a monthUS$0.22 per 1,000 callsPro: US$149 a month for 750,000 calls, plus 250,000 calls at US$0.28 per 1,000. | US$10,025 a monthUS$10.03 per 1,000 calls$25 + 1,000,000 tool calls at $0.01. Auth events not included.Source: arcade.dev/pricing, read 22 October 2026 |
Caveats
- Each auth event adds $0.10; the count depends on how often users sign in.
- Enterprise has annual bundles and discounted rates that are not published.
How we priced it
- List prices on the vendor's own pricing page on the date shown, annual billing where the vendor offers it.
- The smallest plan or ladder step that covers the volume, plus published overage where that is cheaper.
- Each vendor bills a different unit. The unit is named on every row, and the conversion to one action or one tool call is stated.
- Where no public price covers the volume, the cell says so. Nothing is estimated from a sales quote.
actionlake's figures are the USD list. Searching the catalogue, checking a call, planning, cached answers and failed calls are not billed. See every plan on the pricing page.
Moving over
Migration notes
- If each action must run as the individual end user through your identity provider, Arcade is built for that; actionlake connections belong to the tenant.
- For agents that work across your organization's systems, connect each system to actionlake once and give each agent a scoped key.
- Replace gateway tool lists with key scopes: allowed actions, a monthly cap, and approval for writes.
- Keep custom Arcade tools where they hold logic; an actionlake connector can be drafted from an API's documentation.
Questions
Asked about Arcade and actionlake
How is Arcade priced?
Free includes 2,000 tool calls and 2,000 auth events a month. Team is $25 a month plus $0.01 per tool call and $0.10 per auth event. Enterprise is custom, with VPC or air-gapped deployment (arcade.dev/pricing, checked 22 October 2026).
Does actionlake act as the individual user like Arcade?
No. actionlake connections belong to the tenant, and each agent gets a key with its own scope, cap and approval rules. Arcade is designed for each action to run as the signed-in user.
Does Arcade have human approvals?
The Arcade pages we read describe policy hooks that can block or redact before and after a tool call, not a step where a person approves. actionlake can hold a write until a named person approves it.
Can I bring my own MCP server?
Arcade gateways federate custom MCP servers. In actionlake an agent can draft a connector from an API's documentation, test it and sign it.
Sources
- Arcadearcade.dev · read 22 October 2026
- Arcade pricingarcade.dev/pricing · read 22 October 2026
- Arcade MCP gatewaysdocs.arcade.dev/en/guides/mcp-gateways · read 22 October 2026
We wrote these pages, so read them with that in mind. Each competitor is stronger somewhere, and each page says where. If a fact here is out of date, tell us at [email protected] and we will correct it.
Give your agent every system, with limits.
Start free - 100,000 calls/month, no card. No signup. Hard caps on every key. Cached answers are free.