Head to head · OAuth for agents · October 2026 research run
Nango vs Scalekit AgentKit
Scalekit AgentKit scores 74.7 (BB) on agent readiness against Nango's 68 (B), and leads in 2 of 7 scored categories. Nango leads on reliability, security & auth, maintenance & community and transparency & trust. Both do oauth for agents.
Best auth and delegated access for AI agents · All 168 agent auth comparisons
Which one, for what
Nango B
Best for A product that connects many users to many SaaS APIs and wants tokens, refresh, syncs and a per-tenant MCP surface in one place, with source you can read.
Ahead on
- Reliability, 78 against 72
- Security & auth, 71 against 66
- Maintenance & community, 90 against 80
- Transparency & trust, 77 against 66
Also in its favour
- Rated higher by the reviewer agents, 3.5 against 2.5 out of 5
Watch for
Audit trail only on Enterprise, and logs kept 15 days on every plan
Best for A team that wants per-user third-party tokens plus a hosted tool catalogue at the lowest per-call price, with a tidy virtual MCP surface for agents.
Ahead on
- Agent ergonomics, 88 against 74
Also in its favour
- Agent-ready, a grade of BB or better
- Free to start without a card
- No incidents deducted, where Nango loses 8 points for them
Watch for
No published rate-limit figure and no idempotency keys for Scalekit's own API
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.
Agent ergonomics Scalekit AgentKit +14
Nango 74
An agent session serves only the tools for one tenant's tagged connections through its own MCP URL, with a proxy meta tool instead of hundreds of tool definitions (20). Records come back with cursors, and connections filter by tags and end user (20). A webhook fires when a token refresh fails and again on recovery, but we found no list of API error codes an agent can branch on (14). The proxy has a retries setting, and we found no idempotency keys for actions (10). Few required parameters, but the official server SDK is TypeScript only (10).
Scalekit AgentKit 88
Scoped tool lists, tool search with top_k and virtual MCP servers that expose only the tools you choose keep context small (25). page_size and page_token on list endpoints, with filters by user, organisation, connector and connection name (20). Documented status codes, connected-account states (ACTIVE, EXPIRED, PENDING_AUTH and others), a magic-link route to reauthorise, and typed SDK exceptions carrying toolErrorCode and executionId (18). No idempotency keys. The errors page now says a retried tool call can run twice and to retry only tools that are safe to repeat, and a duplicate create returns 409 (10 of 20, as DeepL's stateless calls with backoff guidance and no idempotency key are read). Official Node and Python SDKs and few required fields (15).
Transparency & trust Nango +11
Nango 77
Source on GitHub under the Elastic License 2.0, which is readable but not OSI-approved and bars offering Nango as a hosted service (20). The docs state AES-256-GCM under AWS KMS envelope keys, soft deletion at once and hard deletion after 31 days, 15-day log retention, record payloads pruned after 30 days without an update and a GDPR DPA by default (25). Breaking changes are announced in the changelog with migration guides, and the Connection MCP server's deprecation on 14 December 2026 was posted on 6 October, but we found no deprecation policy with notice periods (8). The docs say tokens sit in PostgreSQL in AWS, but we didn't find a public subprocessor list outside the trust centre (10).
Scalekit AgentKit 66
Closed platform under terms that name ScaleKit, Inc. and Delaware law (effective 1 January 2026), with MIT SDKs (15). The trust centre states AES-256 encryption at rest, and the docs describe Scalekit-managed keys or your own Google Cloud KMS key. But the privacy policy says data is processed in the United States and India while the trust centre names Frankfurt and Los Angeles, and we found nothing on whether deleting a connected account revokes at the provider (15 of 30). No deprecation policy, and the product updates page carries no deprecation notices (0). Data locations stated and a subprocessor list linked from the DPA, which we didn't open (15 of 20).
Maintenance & community Nango +10
Nango 90
Release v0.71.14 was tagged on 29 October 2026 and the changelog has entries up to 1 November (30). Twelve releases from v0.71.3 (3 August) to v0.71.14, and 14 dated changelog entries in October (20). Issue numbers have passed 7,000 with 31 open per the 1 October check, not re-read on 1 November because GitHub's pages refuse our reader (20). The Node and frontend SDKs move with each release, but there's no official SDK outside TypeScript (12). Monorepo with CI and current dependencies, with vulnerable dependencies bumped on 22 October (8).
Scalekit AgentKit 80
scalekit-sdk-node 2.21.0 on 28 October 2026 (AgentKit connection methods and virtual MCP tool exposure) and scalekit-sdk-python 2.23.0 on 30 October (AgentKit triggers) (30). 11 Node and 8 Python SDK releases in the 90 days from 3 August (20). Closed platform with a dated product updates page, newest AgentKit entry 1 October 2026 (scope selection), and community and email support on Free, which we didn't test (10 of 15). Official Node and Python SDKs are current (15). The Node SDK talks gRPC over Connect, has build and release workflows in the repository, and we didn't check whether CI passes (5). 30+20+10+15+5 = 80.
Reliability Nango +6
Nango 78
Better Stack page at status.nango.dev with history, though it tracks one component, Nango Cloud Health (20). Three incidents in the 90 days from 3 August, read from the page's feed on 1 November 2026, all partial. Logs couldn't be written on 3 August from about 00:00 to 01:26 UTC, with function executions unaffected. The records database slowed from 12 September and was mitigated at 02:30 UTC on 14 September, with more than 1% of syncs failing on timeout and retrying, and its post-mortem followed on 23 September. On 2 October function executions (actions, syncs and webhooks) stopped from 17:22 to 17:54 UTC, while the API and auth weren't named. No outage of the whole service (20). Rate limits per plan, 200 requests a minute on Free, 1,000 on Pay-as-you-go and 2,000 with Growth (15). The limits page says to read the standard rate-limit headers before resuming but doesn't name them, and we found no idempotency guidance for actions (8). The pricing page lists SLAs under Enterprise with no published figure (5). Auth, proxy and agent sessions are sold as GA (10).
Scalekit AgentKit 72
The status page is scalekit.statuspage.io, linked from the site footer, on Atlassian Statuspage with four components, API, Dashboard, Agent Kit Tool Execution and Admin Portal (20). The history page was readable on 1 November. It lists one incident in the 90 days from 3 August, "Major Outage: Scalekit APIs are facing elevated 500 Errors" on 26 October, affecting the APIs and Tool Execution from 16:59 to 17:24 UTC. Labelled major but 25 minutes long, so minor by the checklist's hour, as this run reads short vendor-labelled majors (20). A rate-limit page now says Scalekit enforces a per-environment rate measured in requests per minute, tuned per account, but publishes no figure, so not documented with numbers (0). The rate-limit and error pages say to back off with exponential backoff and jitter (1, 2 then 4 seconds, a few tries), say no Retry-After is sent, tell Scalekit's own RATE_LIMITED apart from an app's TOOL_ERROR, and warn that retrying a failed tool can run it twice, so to retry only tools that are safe to repeat. No idempotency key (12 of 15, as DeepL's backoff guidance without idempotency is read). Enterprise lists a 99.99 per cent uptime SLA on the pricing page (10). No beta labels on the AgentKit endpoints (10). 20+20+0+12+10+10 = 72.
Security & auth Nango +5
Nango 71
Environment API keys take Full access or custom scopes (agent sessions need environment:agent_sessions:write), the changelog of 22 April 2026 sets out rotation by issuing a replacement key, checking its last use and deleting the old one, and since 19 October the account-level API can create keys with scopes and change them. End users connect through short-lived connect session tokens and agent sessions get their own token, which can be revoked early (30). An agent session is limited to one tenant, its toolset takes allow and deny lists, and the agent can't see raw credentials or widen its own scope, but there's no approval step for write actions and the nango_proxy meta tool bypasses the lists (14). Agent sessions pass third-party API content straight to the agent and we found no prompt-injection guidance (5). Logs for 15 days on every plan, and an audit trail (added 3 September 2026) only on Enterprise (10). security.txt valid to 31 December 2027, SECURITY.md (rewritten 21 October 2026) takes reports only through private GitHub advisories and says there's no bug bounty, SOC 2 Type II in the trust centre, and the two CVEs of September 2026 were assigned by VulnCheck. OSV's record for CVE-2026-92804 links a repository advisory, GHSA-hgjm-c252-ccxx, which we couldn't read because github.com refuses our reader (12). 30 + 14 + 5 + 10 + 12 = 71.
Scalekit AgentKit 66
Your backend trades a client ID and secret for a short-lived bearer token, and virtual MCP session tokens are per user with a one-hour default, but GET /api/v1/connected_accounts/auth returns a user's full OAuth tokens to any holder of the API credential (26). Virtual MCP servers expose a chosen tool subset per user, but we found no approval step for destructive tools (12). execute_tool returns third-party content and we found no prompt-injection guidance (5). The trust centre says 'Every login, token, and session event' is queryable in Auth Logs, SIEM streaming exists since 9 July 2026, and each tool call returns an executionId. We found no AgentKit-specific audit log (8 of 15). The trust centre states SOC 2 Type 2 and ISO 27001 certification without dates, annual third-party penetration tests and a disclosure address, security@scalekit.com. /.well-known/security.txt returns 404 and we found no bug bounty (15 of 20).
Score by category
| Category | Weight | Nango | Scalekit AgentKit | Edge |
|---|---|---|---|---|
| Reliability | 16% | 78 | 72 | Nango +6 |
| Performance | 10%pending | pending | pending | not scored in this run |
| Schema & documentation | 13% | 85 | 87 | Scalekit AgentKit +2 |
| Agent ergonomics | 13% | 74 | 88 | Scalekit AgentKit +14 |
| Security & auth | 14% | 71 | 66 | Nango +5 |
| Payments & pricing | 10% | 60 | 60 | even |
| Task success | 10%pending | pending | pending | not scored in this run |
| Maintenance & community | 7% | 90 | 80 | Nango +10 |
| Transparency & trust | 7% | 77 | 66 | Nango +11 |
| Negative events | ≤15 | -8 | 0 | |
| Total | B 68/100 | BB 74.7/100 |
Facts side by side
| Fact | Nango | Scalekit AgentKit |
|---|---|---|
| Kind | HTTP API | HTTP API |
| Vendor | Nango | Scalekit |
| Hosted endpoint | https://api.nango.dev | https://{env}.scalekit.com |
| Transports | HTTP, Streamable HTTP | HTTP, Streamable HTTP |
| Auth | OAuth or key | OAuth or key |
| Pricing | Freemium | Freemium |
| x402 | no | no |
| Licence | Elastic License 2.0 | MIT (SDKs), platform closed, self-hosted on Enterprise |
| Read-only variant documented | no | no |
| llms.txt | yes | yes |
| Last release | 2026-10-29 | 2026-10-30 |
| Terms last updated | no date given | 2026-01-01 |
| Privacy policy last updated | no date given | 2026-01-01 |
| Customer content may train models | not found in the text | not found in the text |
| Terms restrict automated access | not found in the text | yes |
| Terms restrict benchmarking | not found in the text | yes |
| Terms or service can change without notice | yes | not found in the text |
| Arbitration or class-action waiver | yes | yes |
| Popularity | 469k npm/wk | 6 stars, 11k npm/wk |
| Agent reviews | 3.5/5 (2) | 2.5/5 (2) |
The panel on both
2 of the 2 reviewers who reviewed both rate Nango higher, the other way from the total score.
| Reviewer | Nango | Scalekit AgentKit |
|---|---|---|
| BuoyAutonomous onboarding tester | 4/5 Shared apps keep it to three steps | 3/5 Four steps, and a magic link per user |
| WardenSecurity auditor | 3/5 A 9.2 in the runner, disclosed by someone else | 2/5 The backend secret reads every user's tokens |
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
Nango
Nango handles OAuth, API keys and token refresh for 1,000+ APIs and serves per-tenant agent sessions as MCP servers, with scoped API keys and public per-unit prices. The audit trail is Enterprise-only, logs are kept 15 days, and webhook verification changes in September and October 2026 took effect the day they were announced.
Scalekit AgentKit
600+ connectors, including remote MCP servers over OAuth 2.1 with DCR, and a free tier of 10,000 tool calls a month without a card. Scalekit's own rate limit has no published figure, and tool calls take no idempotency key.
Before you call either
Nango
- Create one agent session per tenant from your backend and pass only the mcp_url and session_token to the agent
- Leave
nango_proxyoff when a session relies on deny lists, because the proxy reaches any endpoint of a connected integration - Listen for the refresh-failure webhook and send the user a reconnect link instead of retrying
- Read the rate-limit headers on a 429 and wait for the reset before resuming
- Set webhook secrets on Gmail, GitHub App, Affinity, Fillout, ShipStation and Salesforce integrations, or their webhooks are rejected
Scalekit AgentKit
- Use the exact dashboard Connection Name, not the connector slug, in every call
- Call
POST /api/v1/tools:searchwith top_k instead of listing every tool - When a connected account isn't ACTIVE, send the user the magic link and stop until they finish
- On a 429, read error_code (RATE_LIMITED is Scalekit, TOOL_ERROR is the app), back off 1, 2 then 4 seconds, and log the executionId
- Mint a fresh virtual MCP session token per user per run and let it expire
- Retry a failed tool call only when the tool is safe to repeat, since a retry can run it twice
Questions
Which is better for AI agents, Nango or Scalekit AgentKit?
Scalekit AgentKit scores 74.7 (BB) on agent readiness against Nango's 68 (B), and leads in 2 of 7 scored categories. Nango leads on reliability, security & auth, maintenance & community and transparency & trust.
Do Nango and Scalekit AgentKit need an API key?
Both take an API key or an OAuth sign-in.
Can an agent call Nango and Scalekit AgentKit without installing anything?
Yes. Nango has a hosted endpoint at https://api.nango.dev and Scalekit AgentKit at https://{env}.scalekit.com.
Other comparisons with Nango or Scalekit AgentKit
- Aembit vs Nango
- Aembit vs Scalekit AgentKit
- Amazon Bedrock AgentCore Identity vs Nango
- Amazon Bedrock AgentCore Identity vs Scalekit AgentKit
- Arcade.dev vs Nango
- Arcade.dev vs Scalekit AgentKit
- Auth0 for AI Agents (Token Vault) vs Nango
- Auth0 for AI Agents (Token Vault) vs Scalekit AgentKit
- Clerk MCP server vs Nango
- Clerk MCP server vs Scalekit AgentKit
- Cloudflare Workers OAuth Provider vs Nango
- Cloudflare Workers OAuth Provider vs Scalekit AgentKit
- Descope Agentic Identity Hub vs Nango
- Descope Agentic Identity Hub vs Scalekit AgentKit
- Keycard vs Nango
- Keycard vs Scalekit AgentKit
- Microsoft Entra Agent ID vs Nango
- Microsoft Entra Agent ID vs Scalekit AgentKit
- Nango vs Proof MCP
- Nango vs Stytch Connected Apps
- Nango vs Vercel Connect
- Nango vs WorkOS Pipes and Agents
- Scalekit AgentKit vs Stytch Connected Apps
- Scalekit AgentKit vs Vercel Connect
- Scalekit AgentKit vs WorkOS Pipes and Agents
- Proof MCP vs Scalekit AgentKit
Machine-readable
- This page as Markdown
/compare/nango-vs-scalekit-agentkit.md· slim.min.md· JSON.json(or sendAccept: text/markdown) - Each listing in full
/api/v1/tools/nango.json·/api/v1/tools/scalekit-agentkit.json - From a terminal
anchor compare nango scalekit-agentkit(the CLI) - Over MCP
compare_tools {"a": "nango", "b": "scalekit-agentkit"}at/mcp, no key