novyx-mcp

novyx-mcp provides a Model Context Protocol (MCP) server for persistent agent memory. It supports local operation using SQLite for core memory actions, and can connect to Novyx Cloud for an expanded feature set including audit/rollback, knowledge graph, shared context spaces, replay/tracing, evaluation, and governed/autodefense/threat-intel-related tools (depending on plan).

Evaluated Mar 30, 2026 (0d ago)
Homepage ↗ Repo ↗ Ai Ml mcp agents persistent-memory audit-trail rollback knowledge-graph python
⚙ Agent Friendliness
61
/ 100
Can an agent use this?
🔒 Security
60
/ 100
Is it safe for agents?
⚡ Reliability
28
/ 100
Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality
70
Documentation
75
Error Messages
0
Auth Simplicity
85
Rate Limits
25

🔒 Security

TLS Enforcement
80
Auth Strength
55
Scope Granularity
40
Dep. Hygiene
55
Secret Handling
70

Security posture is partially inferable: communication is presumably over TLS when using Novyx Cloud, and the README indicates the API key is supplied via environment variable. However, the documentation does not describe token scopes, least-privilege, key rotation, transport/security details, or how secrets are handled/logged. Dependency hygiene cannot be assessed from the provided README/manifest alone (only dependency names/version constraints are shown).

⚡ Reliability

Uptime/SLA
20
Version Stability
35
Breaking Changes
30
Error Recovery
25
AF Security Reliability

Best When

You want persistent, multi-session memory integrated into MCP-capable clients (Claude Desktop/Cursor/Claude Code) with optional cloud-backed governance and audit features.

Avoid When

You need a public HTTP/SDK-style API with OpenAPI/SDK contracts, or you need explicit, documented rate limits/error codes/retry/idempotency guarantees in the client contract.

Use Cases

  • Give multiple AI agents shared, persistent memory across sessions
  • Maintain an audit trail and support rollback of memory changes
  • Build and query a knowledge graph of facts and relationships
  • Support time-travel debugging via replay/diff of memory state
  • Gate or approve sensitive actions through governed workflows
  • Improve memory quality over time with eval baselines/drift checks
  • Operationalize threat intelligence and auto-defense workflows via MCP tools (plan-dependent)

Not For

  • Use as a general-purpose database/REST service for arbitrary apps (it is an MCP tool server)
  • Environments requiring strictly documented API error codes/retry semantics at the MCP layer (not provided in README)
  • Use where strong compliance/data residency guarantees must be contractually verified via documentation (not shown here)

Interface

REST API
No
GraphQL
No
gRPC
No
MCP Server
Yes
SDK
No
Webhooks
No

Authentication

Methods: NOVYX_API_KEY via environment variable (NOVYX_API_KEY)
OAuth: No Scopes: No

Authentication appears to be via a single API key set in the MCP server process environment. The README does not describe fine-grained scopes, token scopes, or expiration/rotation.

Pricing

Free tier: Yes
Requires CC: No

Pricing is only lightly described. The README indicates some tools (e.g., Replay/Cortex/Threat Intel/Auto Defense/Correlation/Pro+) are plan-dependent, but does not map tool access to specific paid tiers.

Agent Metadata

Pagination
none
Idempotent
False
Retry Guidance
Not documented

Known Gotchas

  • Local mode omits the API key (SQLite only), but cloud mode requires NOVYX_API_KEY; tool behavior and available features differ by mode/plan.
  • Many tools suggest multi-step/approval workflows (e.g., governed actions, drafts/branches); agents should manage state transitions explicitly rather than assuming immediate finalization.
  • The README lists many tools, but does not document parameter schemas, rate limits, or error/retry semantics for each tool.

Alternatives

Full Evaluation Report

Comprehensive deep-dive: security analysis, reliability audit, agent experience review, cost modeling, competitive positioning, and improvement roadmap for novyx-mcp.

AI-powered analysis · PDF + markdown · Delivered within 30 minutes

$99

Package Brief

Quick verdict, integration guide, cost projections, gotchas with workarounds, and alternatives comparison.

Delivered within 10 minutes

$3

Score Monitoring

Get alerted when this package's AF, security, or reliability scores change significantly. Stay ahead of regressions.

Continuous monitoring

$3/mo

Scores are editorial opinions as of 2026-03-30.

6533
Packages Evaluated
19870
Need Evaluation
586
Need Re-evaluation
Community Powered