Navige vs Overmind

Navige 42

No OpenAPI or other machine-readable contract. /openapi.json and /openapi return 404 and the docs link none. The MCP endpoint forwards tools from servers the customer registers and defines none of its own (0). No llms.txt on navige.ai or app.navige.ai and no Markdown versions of the docs (0). The docs say what each route does and does not do, for example that /usage gates nothing, that Watchtower changes no traffic and that error is not deny (15 of 20). /enforce and /usage have field tables with types, a required tool, a classification enum and 200-character bounds that reject rather than truncate. arguments is a free-form object (10 of 15). curl, Python and TypeScript examples on every route, the seven decision shapes, an HTTP status table, five degraded-decision cause codes and a troubleshooting table (14 of 15). No changelog, and /enforce, /usage and /mcp are unversioned paths beside /v1/transactions (3 of 15).

Overmind 81

The docs say each deployment serves its OpenAPI schema at /api/schema/, and the repository holds a generated client of 451 files under frontend/src/openapi. api.overmindlab.ai's robots.txt allows only /.well-known/, so the schema itself wasn't read. Every MCP tool takes a Pydantic input model with extra="forbid", compiled to JSON Schema in overbae/services/mcp/catalog.py, read in the source at main on 31 October 2026 (25). llms.txt with 42 pages, llms-full.txt and Markdown for every page on Mintlify (10). Tool descriptions say what each does and what it doesn't ("Does not activate application routing", "no execution", "Passive: never starts provider work"), and the MCP reference warns that an inspection-style name doesn't mean zero cost. Some workbench descriptions are clipped to the point of being hard to parse (16). Typed inputs with Literal enums, length and range bounds, UUID formats and required fields (13). The docs have worked examples for the CLI, SDK, OTLP and inference, and document status classes (401, 402, 403, 404, 409) and structured MCP errors with code, message, retryability and field details, without a full error-code list (10). The MCP server exposes a contract version and a catalogue fingerprint at overmind://interface/current. The SDK changelog has one undated Unreleased section and no product changelog was found (7). 25+10+16+13+10+7 = 81.

Navige 51

No dated changelog. The latest release found is navige 3.0.0 on npm on 7 September 2026, 40 days before this check (20). navige 2.0.0 (26 July), 2.0.1 (12 August) and 3.0.0 (7 September) fall within 90 days, as do three releases of n8n-nodes-navige (20). A closed service with no public changelog. Support is hello@navige.ai, with a note on each docs page that a person reads reports. The SDK repository's issues weren't readable through the GitHub API from this session (3 of 15). Not in the official MCP registry under either name. The npm SDK covers intercept() and approval status only, with 15 methods that now throw, and the Python SDK couldn't be checked (5 of 15). The SDK has no dependencies, and its repository has no CI or tests (3 of 10).

Overmind 76

overmind 0.1.115 on PyPI on 2026-10-31, and the last commit on main the same day (30). 41 PyPI releases since 21 September 2026 and 24 git tags from v0.0.15 on 18 September to v0.1.23 on 30 October (20). github.com answered 403 to this session, so issues and pull request replies weren't read. 172 commits on main from eight people since 15 September, a Discord server and a support email (8). One official SDK, Python, current with the platform (10). Public CI workflows for the platform and SDK, Dependabot bumps merged on 28 October and a CI badge. We couldn't see the pass state (8). 30+20+8+10+8 = 76.

Navige 53

One nv_ key per agent, stored as a SHA-256 hash, up to ten per agent, each with last use shown, an optional expiry and revocation that the docs say may take a minute to apply. Keys have no scopes beyond the agent. 25, less 10 because the docs give ?api_key= in the URL as the MCP route and as the proxy fallback for clients without custom headers (15 of 30). Deny policies, approval holds and autonomy levels L1 to L5 on paid plans, with write, destructive and financial calls failing closed and a kill switch on every tier. On Watchtower, and on /enforce generally, the decision is advisory and the caller acts on it (15 of 20). Tool results from registered MCP servers return through Navige to the client. We found no prompt-injection mitigation or guidance (3 of 15). Every call is recorded per agent in an append-only ledger, with signed receipts and optional Polygon anchoring (15). A disclosure policy with a two-business-day acknowledgement and safe harbour, but the security.txt it points to returns 404 on all three hosts, there is no bounty, and DPA 1.2 says Navige holds no certification and has not begun a formal process (5 of 20).

Overmind 70

API keys are account-scoped or pinned to one project, carry read, write or both, expire after 30 to 365 days or never, are shown once and show last use, and rotation is documented as replace then revoke. MCP also takes OAuth 2.1 from api.overmindlab.ai with PKCE S256, dynamic registration, a revocation endpoint, one-hour access tokens and rotating refresh tokens, with scopes overmind:read and overmind:write, the default being read. A key can only mint keys with the scope and permissions it holds, since 29 October 2026 (30). Read keys see only read tools over MCP, and the REST API refuses writes from a read-only key since the same commit (#263). Only two tools are destructive and lifecycle tools are allowlisted by name in the catalogue. The docs tell agents to confirm the GPU and rate with the user before starting a dedicated deployment, but there's no server-side approval step (15). The MCP prompts tell the agent to treat code, comments and data as untrusted content, and script output is labelled untrusted in resources. We found no guidance for agents reading trace inputs and outputs back (8). The Console's notification list shows what agents and the CLI did in a project, and overmind://activity returns it over MCP. No exportable audit log was found (10). SECURITY.md lists GitHub private vulnerability reporting and an email with acknowledgement within 3 business days. No security.txt on either host and no bug bounty. The terms claim an ISO 27001 certified management system and the site links a Vanta Trust Center, whose robots.txt disallows crawling, so neither was checked (7). 30+15+8+10+7 = 70.

Navige 61

The proxy returns the provider's response unmodified, decision responses are a few fields, and result capping, tool scoping and context trimming exist to shrink what reaches the agent's context, though all are off by default and paid only (15 of 25). The public API has no list endpoints, since logs and rules moved to the dashboard, and the output ceiling and result caps are the only size controls (8 of 20). Errors are distinct and recoverable, with four 401 reasons, 403 for an inactive or expired agent or the kill switch, 404 for an unknown provider, 429 for a budget and unavailable: true on outages (18 of 20). x-navige-transaction makes retried model calls count once and wait: false avoids a five-minute hold, but /enforce and /usage take no idempotency key. Navige adds no MCP tools of its own to annotate (10 of 20). One base URL and one header, with only tool required on /enforce. One official SDK, for Node.js, and the Python SDK named on the security page couldn't be checked (10 of 15).

Overmind 76

About 85 MCP tools in the source, all loaded for a read-write connection. A key or OAuth grant without write sees only the read tools (about 22), which is the only subset (5, plus 5 for the read-only filter). Cursor and offset pagination, limits by item count and response bytes, preview_rows bounded 1 to 1,000, query filters on traces by status, duration, tokens, cost and time, and evidence_resource links for paging retained results (20). MCP errors carry a structured code, message, retryability and field details, and authentication failures are kept apart from workflow failures. REST validation errors include details (18). Stable request_key on paid and long-running operations with conflict on a changed payload, and every tool sets readOnlyHint, destructiveHint, idempotentHint and openWorldHint from its catalogue entry. Only two tools are marked destructive (20). Account connections must pass project_id on every project tool. One official SDK, Python, plus OpenTelemetry from any language and the OpenAI client for inference (8). 10+20+18+20+8 = 76.

Navige 50

A closed service with MIT SDKs. The Terms (version 1.0, 13 October 2026) leave prices, liability, warranties, service levels, notice periods and governing law marked TBD (5 of 30). DPA 1.2 (14 October 2026) is specific and corrects version 1.1 in public, stating that no automatic deletion runs, that prompt text through the proxy is not stored and that production data is in the United States. The privacy policy (6 October 2026) contradicts it with per-plan retention of 7 days to 1 year on plans the DPA says no longer exist, and speaks of storing masked prompts (12 of 30). No deprecation policy. The SDK 3.0.0 README lists what was removed, and the DPA promises 30 days' notice of material changes and new deletion schedules, while the Terms leave notice periods TBD (4 of 20). DPA Schedule 2 names Railway (United States), Resend, Stripe and Framer (Netherlands), with transfer safeguards. Google, whose analytics tag the website loads, isn't listed, and locations aren't given for Resend or Stripe (14 of 20).

Overmind 65

Platform under AGPL-3.0 and SDK and CLI under MIT, both OSI licences, with a paid commercial licence as an option and a licensing page that says which code is which (30). The privacy policy of 3 October 2026 gives retention (subscription plus 30 days, deletion within 90 days, logs 365 days, backups 7 days), AWS eu-west-1 hosting and OAuth token lifetimes. The terms of 1 March 2026 still describe an earlier security product, say data stays in UK and EU data centres while the privacy policy names model, training and inference providers without locations, and incorporate a DPA that isn't published. The two documents agree on retention but not on the product (15). The terms promise 90 days' written notice before withdrawing a capability and 30 days' notice of material changes. No dated deprecation notices were found (8). SDK analytics are an opt-in extra with three documented off switches. The hosted MCP server reports to PostHog with payloads dropped, which the docs don't mention, and no named sub-processor list was found beyond AWS, though the docs name Modal, OpenRouter, Clerk and Stripe (12). 30+15+8+12 = 65.

Navige 28

Graded as a hosted service. A public Better Stack page at status.navige.ai exists, but its one monitor is api.trustloop.live/health, the host of the product's former name, and not api.navige.ai (5 of 20). That monitor shows downtime every day from 26 July to 17 October 2026 (7.221 per cent uptime) under a header saying all services are online, so there is no readable history for the current API (5). No rate-limit numbers in the docs, onboarding guide or terms (0). The docs separate a 503 outage from a 400 bug, state fail-open and fail-closed defaults per action class, return 429 for a blocking budget, pass provider errors through for the client's own retry logic, and make retries count once under x-navige-transaction. No Retry-After or backoff guidance (8 of 15). The Terms leave service levels as TBD (0). No beta or preview label on the gateway (10).

Overmind 23

No status page. status.overmindlab.ai doesn't resolve, /status returns 404 and neither the site nor the docs link one (0). With no history to read, 5. No request rate limits for the REST API, MCP or OTLP ingest were found in the docs. The only pacing settings, JEV_REQUESTS_PER_MINUTE and JEV_TOKENS_PER_SECOND, are self-hosting configuration for the scoring model (0). No 429 or Retry-After guidance, but the docs give safe-retry rules in detail. Launches, uploads, pipeline runs and inference take a stable request_key, identical retries return the existing receipt, changed payloads conflict, and run_evaluation is named as not idempotent (8). The terms promise commercially reasonable efforts toward 99.5 per cent availability a month, with remedies only in an order form or separate SLA, so half (5). The site doesn't call the service beta, but the SDK is classed "3 - Alpha" on its package metadata and the public repository is seven weeks old (5). 0+5+0+8+5+5 = 23.