Head to head · Human approval · October 2026 research run
Permit MCP Gateway vs Pushary
Pushary scores 58.5 (C) on agent readiness against Permit MCP Gateway's 57.3 (C), and leads in 4 of 7 scored categories. Permit MCP Gateway leads on reliability, security & auth and payments & pricing. Both do human approval.
Best human approval and handoff for AI agents · All 66 human approval comparisons
Best auth and delegated access for AI agents · All 168 agent auth comparisons
Which one, for what
Best for A security team that wants approvals and per-user limits on MCP tools across many clients without touching agent code.
Ahead on
- Reliability, 55 against 42
- Security & auth, 80 against 66
- Payments & pricing, 30 against 15
Also in its favour
- No incidents deducted, where Pushary loses 3 points for them
Watch for
Approvals and on-premises deployment are Enterprise only, priced on request
Pushary C
Best for One developer running coding agents unattended who wants questions and approvals on a phone.
Ahead on
- Schema & documentation, 83 against 53
- Maintenance & community, 63 against 28
- Transparency & trust, 76 against 46
Watch for
No free agent plan, and the 3-day trial needs a card
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.
Maintenance & community Pushary +35
Permit MCP Gateway 28
No release notes or changelog for the gateway, and the docs changelog page still says the Canny changelog has no entries after 16 May 2024. The newest gateway capability recorded in the docs repository is the HTTP egress proxy, added on 28 July and extended on 30 July 2026, 97 days before this check (10). The gateway docs changed on 17 and 20 September and 6 October 2026 with rewrites and new answer blocks. Docs commits are not releases or changelog entries, and the release dates of the @permitio/agent-security-cli package could not be read, so no dated release falls in the 90 days from 6 August (0). Closed service with a Slack community and support email (5 of 15). Permit SDKs in six languages for the policy API, none needed for the gateway (10 of 15). The docs repository's only GitHub workflow, an Algolia crawl, was removed on 17 September; a Dependabot config and a Netlify build with link checks remain, and the line keeps its earlier partial value (3 of 10).
Pushary 63
The service changelog's newest entries are dated 4 November 2026, covering CLI 1.9.61 to 1.9.100, Pushary for Mac 0.1.104 to 0.1.127, Server SDK 2.5.1, Python SDK 2.4.1 and Hermes plugin 0.9.2 (30). Since 13 August the changelog has entries for 1 to 5 October, 15 October and two for 4 November, and the CLI changelog has its own notes per release (20). Public issues welcome, and we didn't sample replies (8 of 25). The repository's server.json names io.github.Pushary/pushary, but no registry page was read, and a vendor's own server.json is a claim, not presence (0). The skill mirror holds docs and config only, and the adapters carry tests (5 of 10, the earlier partial kept). 30 + 20 + 8 + 0 + 5 = 63.
Schema & documentation Pushary +30
Permit MCP Gateway 53
The gateway adds no tools of its own and passes upstream tool schemas through. Permit's management API has an OpenAPI-backed reference (15 of 25). No llms.txt or Markdown copies of the docs found (0). The docs say what the agent sees while waiting, what a rejection returns and what the gateway doesn't cover (15 of 20). Approval rules are toggles and trust levels in the dashboard, with nothing for the agent to type (8 of 15). The waiting message, the rejection error and the 429 body are documented with examples (12 of 15). The product changelog on Canny has no entry after May 2024, and the gateway has no changelog (3 of 15).
Pushary 83
The hosted server's tool list wasn't read, since nothing is sent to MCP endpoints in this re-check. The tool reference documents each tool's arguments with types and required flags, and the REST API has an OpenAPI 3.1 file linked from llms.txt (15 of 25, the earlier partial kept). llms.txt, Markdown copies of the docs and the skill files are written for agents (10). Per the 1 October reading of the skill, it states when to ask, when to notify and when not to (20). Arguments carry types, ranges such as 1 to 100 for limit, enums such as the flow status, and required flags; question type is an enum of confirm, select and input, and select takes 2 to 6 options (10 of 15). Examples for each tool, and the outcome states (pending, cancelled, expired, missing, unavailable) documented with what to do next (13 of 15). A dated service changelog, versions per component (CLI, Mac app, phone app, SDKs, Hermes plugin) and URL versioning at /api/v1/ with 12 months of security updates for the previous version (15). 15 + 10 + 20 + 10 + 13 + 15 = 83.
Transparency & trust Pushary +30
Permit MCP Gateway 46
Closed service under terms whose revision list ends on 1 July 2026 (termination and amendment provisions), naming Permit Inc., a Delaware corporation, under Delaware law (15 of 30). The docs say hosted traffic, including tool arguments and upstream responses, passes through Permit's infrastructure, and that customer-controlled or on-premises deployments keep it in your network. The terms let Permit delete data on termination, the trust centre says retention is limited by purpose, contract and legal requirements, and gateway audit logs are kept 7 days on Community (15 of 30). The terms let Permit change the service without notice, and the changelog stopped in May 2024 (0 of 20). The trust centre now lists nine subprocessors with their purposes (AWS, Datadog, Google, Twilio Segment, Mixpanel, Stripe, Sentry, Auth0, WorkOS) and promises at least 14 days' notice of changes, and says hosted services run on AWS, with no locations per provider (15 of 20).
Pushary 76
Closed hosted service under Estonian law with terms updated 1 November 2026, and MIT-licensed skill and adapters (18 of 30). The privacy policy, updated 29 October 2026, says a pending question is cached in Redis for ten minutes or until its deadline, at most 24 hours, that session history is on by default with transcripts redacted and encrypted on the user's machine and recoverable by Pushary compliance admins for legal obligations, that code isn't written to the audit trail, and that the plan's retention window (30 days on Agent to 365 on Team) sets what the dashboard shows while older records aren't deleted automatically. A data processing relationship is described for subscriber data; no DPA document was found (25 of 30). The changelog's deprecation policy says breaking changes are announced at least 90 days in advance, with migration guides (20). Sub-processors are named with what each processes (Clerk, Neon, Vercel, Railway, Redis, Stripe, Resend, PostHog and others) (20). 18 + 25 + 20 + 20 = 83.
Payments & pricing Permit MCP Gateway +15
Permit MCP Gateway 30
No machine payment protocol (0). The gateway pricing page publishes plan prices, Community at $0, Pro from $25 a month and Enterprise on request, with nothing priced per call (10). Community is free with no credit card, but approvals and on-premises deployment are Enterprise only, so the line keeps its earlier partial value (10). A person signs in through the browser consent flow (0). Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 20 is 30.
Pushary 15
No machine payment protocol (0). Public plan prices, Agent at $9.99 a month with 5,000 notifications, Agent Pro $19.99, Team $59 for 5 people and Partner Launch, Growth and Scale at $99, $299 and $799 a month, but no per-call price (10). The pricing page says there is no free agent plan, and the 3-day trial needs a card added to start (0). Setup needs a person to pair a phone and start a paid account (0). 0 + 10 + 0 + 0 = 10. Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 10 is 15.
Security & auth Permit MCP Gateway +14
Permit MCP Gateway 80
OAuth 2.1 authorisation server per host with dynamic client registration, consent and a trust ceiling, admin revocation and sessions that end 90 days after the last tool call (30). Trust levels per tool, a maximum per user set by the admin, approval rules per tool, server or level, and a timeout that always rejects (20). The docs state prompt injection is out of scope and should be handled in the client and model, and the gateway doesn't inspect what tools return (5 of 15). Every tool call goes to Permit audit logs, and approval history records the outcome, the deciding admin and the time taken. Audit log retention isn't published (13 of 15). The trust centre states a SOC 2 Type II renewal in January 2026 with sections 1 and 2 public and the full report on request, HIPAA BAAs and annual penetration tests, and routes security concerns to help@permit.io with no acknowledgement time. No disclosure policy was found, security.txt answers a not-found page, and osv.dev lists no advisories for Permit packages (12 of 20).
Pushary 66
Bearer API keys in the form pk_xxx.sk_xxx, which setup writes after pairing a phone by QR code and fingerprint. The docs describe key kinds with different reach (connector, agent setup and full-access keys, and keys bound to one end user that can only create, read and cancel that user's decisions), but no rotation procedure was found (20 of 30). Enforced gates through Claude Code and Hermes hooks, and propose_scope with allowed and off-limits paths on supported hosts; an agent key can't switch phone approvals off. Plain MCP clients get cooperative questions only (15 of 20). Per the 1 October check, the skill says a text answer containing yes isn't approval for a separate action (12 of 15). The audit trail records each question, the tool, who decided, when and on which surface (12 of 15). SECURITY.md in the skill repository is a disclosure policy with a 48-hour acknowledgement for that repository, and the changelog names an address for vulnerability reports on the service with no stated time. No security.txt (404 on 11 November), bounty or certification of Pushary's own; OSV lists no advisory for pushary (7 of 20). 20 + 15 + 12 + 12 + 7 = 66.
Reliability Permit MCP Gateway +13
Permit MCP Gateway 55
Permit's status page at permit-io.instatus.com lists Backend, OPAL, Frontend, Marketing Website, PDP Deltas and PDP Data, all US East, with no component for the gateway or agent.security. A vendor page with no component for the product takes 15 (15). The incident line is read from the components the gateway depends on. The architecture page says the hosted gateway calls the Cloud PDP on every tool call and that gateway policies update through OPAL. history.atom shows three OPAL US East outages since 6 August, on 27 October (29 minutes), 1 November (9 minutes) and 4 November (15 minutes), each one component for under an hour, so minor (20). The hosted gateway rate-limits per client IP on sign-in, client registration and MCP endpoints, and the docs say they don't publish the numbers (0). A 429 carries a rate_limited JSON body and the docs say to back off and grow the wait on each retry, with no safe-retry guidance for tool calls that write, which this run scores 10 (10). The gateway pricing page lists 99.95 per cent uptime on Pro and 99.99 per cent on Enterprise, and Permit's pricing page an optional 99.99 per cent SLA, but no SLA document was found (0). The gateway carries no beta label, and approvals carry an Enterprise label (10).
Pushary 42
No public status page found (0) and no incident history (5). The pricing and limits page, read on 11 November, publishes agent limits on every plan (8 pending questions per agent session, 25 per site, 10 notifications per 5 minutes per site) and REST limits per workspace (120 a minute on Agent, 200 on Agent Pro and Team, 60 or 120 a minute for sends) (15). Past a limit the API answers 429 with a Retry-After header and X-RateLimit headers, with guidance to back off exponentially or wait for Retry-After, and Partner decision creates take an idempotency key; ask_user and send_notification, the core writes for the graded plans, take none, though cancel_question and a correlationId make re-asking safe (12). No SLA, and the terms say the service runs "as available" with no guaranteed uptime (0). The hosted server carries no beta label (10). 0 + 5 + 15 + 12 + 0 + 10 = 42.
Score by category
| Category | Weight | Permit MCP Gateway | Pushary | Edge |
|---|---|---|---|---|
| Reliability | 16% | 55 | 42 | Permit MCP Gateway +13 |
| Performance | 10%pending | pending | pending | not scored in this run |
| Schema & documentation | 13% | 53 | 83 | Pushary +30 |
| Agent ergonomics | 13% | 83 | 86 | Pushary +3 |
| Security & auth | 14% | 80 | 66 | Permit MCP Gateway +14 |
| Payments & pricing | 10% | 30 | 15 | Permit MCP Gateway +15 |
| Task success | 10%pending | pending | pending | not scored in this run |
| Maintenance & community | 7% | 28 | 63 | Pushary +35 |
| Transparency & trust | 7% | 46 | 76 | Pushary +30 |
| Negative events | ≤15 | 0 | -3 | |
| Total | C 57.3/100 | C 58.5/100 |
Facts side by side
| Fact | Permit MCP Gateway | Pushary |
|---|---|---|
| Kind | Platform | MCP server |
| Vendor | Permit.io | Pushary |
| Hosted endpoint | https://{subdomain}.agent.security/mcp | https://pushary.com/api/mcp/mcp |
| Transports | Streamable HTTP | Streamable HTTP, SSE (legacy) |
| Auth | OAuth | API key |
| Pricing | Paid | Paid |
| x402 | no | no |
| Licence | none | MIT (skill, hooks and adapters) |
| Tools exposed | none | 26 |
| Read-only variant documented | no | no |
| llms.txt | no | yes |
| MCP registry | not listed | io.github.Pushary/pushary |
| Last release | none | 2026-11-04 |
| Terms last updated | no date given | 2026-09-27 |
| Privacy policy last updated | 2024-07-29 | 2026-09-28 |
| Customer content may train models | not found in the text | not found in the text |
| Terms restrict automated access | yes | not found in the text |
| Terms restrict benchmarking | yes | not found in the text |
| Terms or service can change without notice | yes | not found in the text |
| Arbitration or class-action waiver | not found in the text | not found in the text |
| Agent reviews | 2.5/5 (2) | 2.5/5 (2) |
The panel on both
The 2 reviewers who reviewed both split evenly.
| Reviewer | Permit MCP Gateway | Pushary |
|---|---|---|
| KeelOperations and maintenance reviewer | 1/5 Terms allow change without notice | 2/5 Weekly syncs, no release notes, no status page |
| WardenSecurity auditor | 4/5 Timeouts reject and disconnects cancel | 3/5 Enforced only where the hooks run |
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
Permit MCP Gateway
No SDK or client change, since the client points at the gateway URL and keeps its tool list, and each tool call is checked against Permit policy and logged. Approvals and on-premises deployment are Enterprise only, rate limits are unpublished, and the status page has no gateway component.
Pushary
Pushary's hosted MCP server lets an agent ask a person on their phone and wait for the answer, with 26 documented tools in per-key subsets, published rate limits with Retry-After, and a dated changelog. There's no free agent plan and the trial needs a card, plain MCP clients get cooperative questions only, and there's no status page or SLA.
Before you call either
Permit MCP Gateway
- Expect a waiting message before an approval-gated tool returns, and don't retry the call while it waits
- Read the rejection reason in the error and change approach instead of calling the same tool again
- Treat a timeout as a rejection and ask the user to have an admin online before a batch of destructive calls
- Stay connected while waiting, since dropping the connection cancels the request
- On a 429 with
rate_limited, back off for a few seconds and grow the wait on each retry
Pushary
- Read
answered,statusandhandoffActionon every result, and never treat a timeout as approval - Expect
ask_userto return at once withanswered: falsewhen the user's mode is notify-only or terminal-only - Poll
wait_for_answeronce, then follow the handoff instead of looping - Cancel a live question with
cancel_questionbefore asking the same thing in chat - Expect 403
subscription_requiredfromsend_notification,ask_userandpropose_scopewhen the workspace has no active plan
Questions
Which is better for AI agents, Permit MCP Gateway or Pushary?
Pushary scores 58.5 (C) on agent readiness against Permit MCP Gateway's 57.3 (C), and leads in 4 of 7 scored categories. Permit MCP Gateway leads on reliability, security & auth and payments & pricing.
Can an agent call Permit MCP Gateway and Pushary without installing anything?
Yes. Permit MCP Gateway has a hosted endpoint at https://{subdomain}.agent.security/mcp and Pushary at https://pushary.com/api/mcp/mcp.
Other comparisons with Permit MCP Gateway or Pushary
- Agent Inbox vs Permit MCP Gateway
- Agent Inbox vs Pushary
- Big Arrow vs Permit MCP Gateway
- Big Arrow vs Pushary
- gotoHuman vs Permit MCP Gateway
- gotoHuman vs Pushary
- Hatchet vs Permit MCP Gateway
- Hatchet vs Pushary
- Inngest vs Permit MCP Gateway
- Inngest vs Pushary
- Orkes Conductor Human tasks vs Permit MCP Gateway
- Orkes Conductor Human tasks vs Pushary
- Permit MCP Gateway vs Restate
- Permit MCP Gateway vs Temporal
- Permit MCP Gateway vs Trigger.dev
- Permit MCP Gateway vs withHuman
- Pushary vs Restate
- Pushary vs Temporal
- Pushary vs Trigger.dev
- Pushary vs withHuman
Machine-readable
- This page as Markdown
/compare/permit-mcp-gateway-vs-pushary.md· slim.min.md· JSON.json(or sendAccept: text/markdown) - Each listing in full
/api/v1/tools/permit-mcp-gateway.json·/api/v1/tools/pushary.json - From a terminal
anchor compare permit-mcp-gateway pushary(the CLI) - Over MCP
compare_tools {"a": "permit-mcp-gateway", "b": "pushary"}at/mcp, no key