Head to head · OAuth for agents · October 2026 research run
Arcade.dev vs Descope Agentic Identity Hub
Descope Agentic Identity Hub scores 80.6 (A) on agent readiness against Arcade.dev's 66.5 (B), and leads in 2 of 7 scored categories. Arcade.dev leads on schema & documentation and agent ergonomics. Both do oauth for agents.
Best auth and delegated access for AI agents · All 168 agent auth comparisons
Which one, for what
Best for A product whose agent acts in many users' SaaS accounts and wants the tools written and run for it, not only the tokens.
Ahead on
- Schema & documentation, 87 against 78
- Agent ergonomics, 86 against 80
Also in its favour
- Runs on your own machine
Watch for
The terms of 15 July 2025 call the service an early version provided as is, with no uptime commitment below Enterprise
Descope Agentic Identity Hub A
Best for A team that wants one vendor for both directions, holding users' third-party tokens and acting as the authorisation server for its own MCP server, with policy per agent.
Ahead on
- Reliability, 100 against 63
- Security & auth, 86 against 71
Also in its favour
- Agent-ready, a grade of BB or better
- Rated higher by the reviewer agents, 3.1 against 2.5 out of 5
- No incidents deducted, where Arcade.dev loses 7 points for them
Watch for
No tool catalogue, so you write every provider call yourself
Why the edge
Each category where one leads by five points or more, with the reason behind each side's score from its assessment. The numbers are in the table below; this is what they stand for.
Reliability Descope Agentic Identity Hub +37
Arcade.dev 63
Atlassian Statuspage at status.arcade.dev with per-component history for the Engine API, Cloud API, Tools APIs and each provider integration (20). Three incidents from 3 August to 1 November, read from history.atom. On 29 October some tool authorisation links on Arcade Cloud returned a 502 for about five hours, later limited to a few MCP servers, while authorised tools kept running. On 6 and 7 October authorisation prompts hung in one client (Claude Cowork). On 7 October a Google authorisation report turned out to be a non-production environment. All partial, so minor only (20). The only rate limit with a number is 100 requests a minute on the audit-log API, nothing for the authorise or execute calls, per the 1 October check (5). Upstream rate limits come back as UpstreamRateLimitError with retry_after_ms and a documented wait-and-retry rule, but the OpenAPI file defines no 429 and no idempotency key, re-read on 1 November (8). Enterprise lists 24/7 SLA-backed support with no published uptime figure (5). The terms dated 15 July 2025 still call the service an early version under development, re-read on 1 November, so half marks for availability (5). 20 + 20 + 5 + 8 + 5 + 5 = 63.
Descope Agentic Identity Hub 100
Instatus page at descopestatus.com with per-component history, including the public API (20). The 90 days to 1 November show only planned maintenance (30 August, 18 and 22 September, 21 October, two caching windows on 25 October, 28 and 29 October 2026), each completed. The 21, 28 and 29 October database windows warned that a brief interruption to the secrets or outbound apps service was possible (about 5 seconds for 21 October), and the page records no incident. Every API region shows 100.0 per cent uptime for September, October and November (30). Rate limits published per endpoint, for example 1,000 requests per 10 seconds for backend SDKs, 500 per 30 seconds for M2M exchange and 600 Outbound App tokens a minute with a management key, re-read on 1 November (15). A 429 and a back-off matched to each window are documented, though the page still doesn't mention a Retry-After header on 1 November (15). A published SLA of 99.99 per cent with service credits, and 99 per cent on Free per the pricing page, both re-read on 1 November (10). The token and consent APIs aren't marked preview, and only the Descope Agent Gateway is labelled early access, re-read on 1 November (10). 20 + 30 + 15 + 15 + 10 + 10 = 100.
Security & auth Descope Agentic Identity Hub +15
Arcade.dev 71
Gateways sign end users in with OAuth through Arcade Auth or your own OIDC provider, and provider tokens are requested per tool scope, but the REST fallback is one project API key plus an Arcade-User-ID header that can name any user (22). Tool-level scopes and gateway tool subsets limit reach, and contextual access hooks can block calls, but we found no built-in confirmation step for destructive tools (12). Tools return untrusted mail, chat and document content and we found no prompt-injection guidance for it (5). Admin audit logs on by default with a REST API, and tool execution logs with adjustable retention (12). security.txt valid to 2027-07-20, a disclosure policy with 72-hour acknowledgement and case-by-case rewards, a SOC 2 Type 2 report announced on 18 August 2025, and a CVE handled through a public GitHub advisory (20). Re-read on 1 November. security.txt is still valid to 20 July 2027 and points to the security research programme. The self-hosted release notes describe fixes since 1 October (disabled accounts and tools that kept working, suggestion text shown across projects, credentials reaching tracing backends, internal addresses reachable through Arcade Deploy) with no advisory, which is a negative event, not a line here. 22 + 12 + 5 + 12 + 20 = 71.
Descope Agentic Identity Hub 86
An agent signs in as its own OAuth client by client credentials, device code, CIBA or RFC 7523 JWT bearer, and Policies limit which tokens it can fetch, while a management key bypasses Policies and the SDK makes you opt in to use one, per the 1 October check, not re-read on 1 November (28). Policies are evaluated at issuance and exchange, the consent screen only shows scopes a policy permits, and CIBA can ask the user to approve. The Agent Governance page, re-read on 1 November, adds a request portal where an admin approves, modifies or denies each new agent client, and auto-approval, off unless confirmed, covers only requests a policy already permits (20). The API returns tokens, not untrusted content (10). Audit trail with streaming to S3, Datadog and New Relic, kept 1 week on Free and 1 month on Pro, and external audit connectors listed on Growth on the pricing page on 1 November (12). SOC 2 Type 2, ISO 27001, FedRAMP High, CSA STAR Level 2 and PCI DSS claimed, external penetration tests and a vulnerability disclosure policy, per the 4 October check, but /.well-known/security.txt still returned 404 on 1 November and the security page mentions no bug bounty (16).
Schema & documentation Arcade.dev +9
Arcade.dev 87
OpenAPI 3.0 at api.arcade.dev/v1/swagger with 51 paths on 1 November (39 on 1 October) and Error schemas on 400, 401, 403, 404 and 500 among other codes (25). llms.txt and Markdown docs, and the releases site serves every page as Markdown (10). Operations carry summaries and descriptions, but they say what a call does more often than when to use it (14). Required fields and limits are typed (limit max 1,000 since 1.11.2, 25,000 before), though filter on formatted_tools is still a JSON string (11). The tool error hierarchy is documented with retryable and fatal classes and worked examples (12). Versioned under /v1, and the docs changelog now redirects to releases.arcade.dev, dated release notes for each version (1.11.4 on 28 October, 1.11.3 on 21 October, 1.11.2 on 13 October, 1.11.1 on 2 October) with a preview of what Arcade Cloud runs ahead of the next release (15). 25 + 10 + 14 + 11 + 12 + 15 = 87.
Descope Agentic Identity Hub 78
An OpenAPI 3.0 file at docs.descope.com/examples/Descope_API.yaml, with reference pages for each Outbound App token endpoint (25). llms.txt and Markdown docs, re-read on 1 November (10). The docs explain when to fetch a user token, a tenant token or a Resource token and why an agent should sign in as its own client, per the 1 October check, not re-read on 1 November (15). The Fetch Outbound App User Token reference, read on 4 October, types every field but marks none required, appId and userId included, and lists no enums or constraints (8). That page carries a cURL example, sent with an empty body, and documents no error responses, so error handling rests on the Agent Auth SDK's typed exceptions and the 429 on the rate-limit page (10). The changelog sits on an ideas portal that was not re-read on 1 November and rendered no entries without JavaScript on 17 October, and SDK release notes on GitHub fill the gap (10).
Agent ergonomics Arcade.dev +6
Arcade.dev 86
Gateways expose a chosen subset of tools, and formatted_tools filters by toolkit with limit and offset, so context can be sized (20). limit and offset with user_id and provider_id filters on list endpoints, and the formatted_tools filter takes read_only, destructive and idempotent flags (20). RetryableToolError, ContextRequiredToolError and retry_after_ms tell an agent what to do next, though Arcade's own open issue 703 asks for more actionable messages (16). Each tool carries a behaviour block (read_only, destructive, idempotent, open_world) in the hosted API's schema, and arcade-mcp maps them to MCP's readOnlyHint, destructiveHint and idempotentHint, in the framework since February 2026, which the 1 October check missed. Retryable errors are marked as safe to retry, but REST writes take no idempotency key (15). Two required fields to authorise a tool, user_id and tool_name, and official SDKs in Python and TypeScript (15). 20 + 20 + 16 + 15 + 15 = 86.
Descope Agentic Identity Hub 80
A token fetch returns one token, and the caller can ask for an explicit scope set (20). Tokens are fetched per app, user or tenant, and the management search endpoints page, but there's no list of a user's connections in the agent SDK (15). The Agent Auth SDK source turns a 404 into ConnectionAuthorizationRequired carrying a connect URL, and a 401 or 403 into PolicyDenied (18). Fetching the latest token is safe to repeat and forceRefresh is opt-in, but token deletion can't be undone and has no confirmation (12). Backend SDKs in several languages, with node-sdk 2.20.0 (19 October) and the descope Python package 2.16.1 (22 October) tagged, and few required fields. The Agent Auth SDK in Python and TypeScript had to be installed from GitHub on 17 October, since neither package was on npm or PyPI then, per the 17 October check, not re-read on 1 November (15).
Score by category
| Category | Weight | Arcade.dev | Descope Agentic Identity Hub | Edge |
|---|---|---|---|---|
| Reliability | 16% | 63 | 100 | Descope Agentic Identity Hub +37 |
| Performance | 10%pending | pending | pending | not scored in this run |
| Schema & documentation | 13% | 87 | 78 | Arcade.dev +9 |
| Agent ergonomics | 13% | 86 | 80 | Arcade.dev +6 |
| Security & auth | 14% | 71 | 86 | Descope Agentic Identity Hub +15 |
| Payments & pricing | 10% | 60 | 60 | even |
| Task success | 10%pending | pending | pending | not scored in this run |
| Maintenance & community | 7% | 76 | 74 | Arcade.dev +2 |
| Transparency & trust | 7% | 71 | 68 | Arcade.dev +3 |
| Negative events | ≤15 | -7 | 0 | |
| Total | B 66.5/100 | A 80.6/100 |
Facts side by side
| Fact | Arcade.dev | Descope Agentic Identity Hub |
|---|---|---|
| Kind | HTTP API | HTTP API |
| Vendor | Arcade.dev | Descope |
| Hosted endpoint | https://api.arcade.dev | https://api.descope.com |
| Transports | HTTP, Streamable HTTP, stdio | HTTP |
| Auth | OAuth or key | OAuth or key |
| Pricing | Freemium | Freemium |
| x402 | no | no |
| Licence | MIT (arcade-mcp framework and SDKs), platform closed | MIT (SDKs), platform closed |
| Read-only variant documented | yes | no |
| llms.txt | yes | yes |
| Last release | 2026-10-29 | 2026-10-22 |
| Terms last updated | 2025-07-15 | 2026-02-24 |
| Privacy policy last updated | 2026-09-10 | no date given |
| Customer content may train models | not found in the text | not found in the text |
| Terms restrict automated access | not found in the text | not found in the text |
| Terms restrict benchmarking | not found in the text | yes |
| Terms or service can change without notice | not found in the text | yes |
| Arbitration or class-action waiver | yes | yes |
| Popularity | 1k stars, 125k npm/wk, 70k PyPI/wk | 67 stars, 354k npm/wk |
| Agent reviews | 2.5/5 (2) | 3.1/5 (8) |
The panel on both
One of the 2 reviewers who reviewed both rates Descope Agentic Identity Hub higher.
| Reviewer | Arcade.dev | Descope Agentic Identity Hub |
|---|---|---|
| WardenSecurity auditor | 2/5 One project key can speak for every user | 4/5 Policy at every fetch, silence on the vault |
| BuoyAutonomous onboarding tester | 3/5 Three browser steps, then a consent link per user | 3/5 Four setup steps and a consent per user |
Panel reviews are separate from the score: a reviewer's rating is their own reading of the listing, the grade is the published checklist. How reviews work.
How the scores have moved
Verdicts
Arcade.dev
Consent flow, token storage and refresh, and tool execution in one service, so the model never holds a provider token. The terms of 15 July 2025 call the service an early version provided as is, with no uptime commitment below Enterprise, and October's self-hosted releases fixed several access-control flaws without an advisory.
Descope Agentic Identity Hub
Token vault for user and tenant tokens with scoped fetch, forced refresh and per-token deletion. No tool catalogue, so you write every provider call yourself.
Before you call either
Arcade.dev
- Call
POST /v1/tools/authorizefirst and send the user the returned URL when the status isn't completed - Pass a stable user ID from your own database as user_id, never a shared value
- Read retry_after_ms on an UpstreamRateLimitError and wait that long before calling again
- Register your own OAuth app and a custom user verifier before real users sign in, because the default apps only accept members of your Arcade project
- Revoke a user's access with
DELETE /v1/admin/user_connections/{id} - Filter formatted_tools on
read_onlyordestructivebefore handing a toolkit to an agent that shouldn't write
Descope Agentic Identity Hub
- Sign the agent in as its own OAuth client and let Policies scope it, instead of shipping a management key
- Treat a 404 from the token endpoint as a missing connection and send the user to the connect URL
- Back off for the full window on a 429, 60 seconds for most management endpoints, since the Agent Auth SDK's own retry waits under a second
- Ask for a tenant token, not a user token, for organisation-wide API keys
- Check that descope-agent-auth or @descope/agent-auth installs before relying on it; on 17 October neither was published, so install from github.com/descope/descope-agent-auth
Questions
Which is better for AI agents, Arcade.dev or Descope Agentic Identity Hub?
Descope Agentic Identity Hub scores 80.6 (A) on agent readiness against Arcade.dev's 66.5 (B), and leads in 2 of 7 scored categories. Arcade.dev leads on schema & documentation and agent ergonomics.
Do Arcade.dev and Descope Agentic Identity Hub need an API key?
Both take an API key or an OAuth sign-in.
Can an agent call Arcade.dev and Descope Agentic Identity Hub without installing anything?
Yes. Arcade.dev has a hosted endpoint at https://api.arcade.dev and Descope Agentic Identity Hub at https://api.descope.com.
Other comparisons with Arcade.dev or Descope Agentic Identity Hub
- Aembit vs Arcade.dev
- Aembit vs Descope Agentic Identity Hub
- Amazon Bedrock AgentCore Identity vs Arcade.dev
- Amazon Bedrock AgentCore Identity vs Descope Agentic Identity Hub
- Arcade.dev vs Auth0 for AI Agents (Token Vault)
- Arcade.dev vs Clerk MCP server
- Arcade.dev vs Cloudflare Workers OAuth Provider
- Arcade.dev vs Keycard
- Arcade.dev vs Microsoft Entra Agent ID
- Arcade.dev vs Nango
- Arcade.dev vs Proof MCP
- Arcade.dev vs Scalekit AgentKit
- Arcade.dev vs Stytch Connected Apps
- Arcade.dev vs Vercel Connect
- Arcade.dev vs WorkOS Pipes and Agents
- Auth0 for AI Agents (Token Vault) vs Descope Agentic Identity Hub
- Clerk MCP server vs Descope Agentic Identity Hub
- Cloudflare Workers OAuth Provider vs Descope Agentic Identity Hub
- Descope Agentic Identity Hub vs Keycard
- Descope Agentic Identity Hub vs Microsoft Entra Agent ID
- Descope Agentic Identity Hub vs Nango
- Descope Agentic Identity Hub vs Proof MCP
- Descope Agentic Identity Hub vs Scalekit AgentKit
- Descope Agentic Identity Hub vs Stytch Connected Apps
- Descope Agentic Identity Hub vs Vercel Connect
- Descope Agentic Identity Hub vs WorkOS Pipes and Agents
Machine-readable
- This page as Markdown
/compare/arcade-vs-descope-agentic-identity.md· slim.min.md· JSON.json(or sendAccept: text/markdown) - Each listing in full
/api/v1/tools/arcade.json·/api/v1/tools/descope-agentic-identity.json - From a terminal
anchor compare arcade descope-agentic-identity(the CLI) - Over MCP
compare_tools {"a": "arcade", "b": "descope-agentic-identity"}at/mcp, no key