Vestige

Vestige is a local-first cognitive memory system for AI agents that implements neuroscience-inspired processes including FSRS-6 spaced repetition, prediction error gating, spreading activation, and memory consolidation (dreaming) — going beyond simple RAG to create a genuinely evolving memory that forgets less-relevant information over time.

Evaluated Mar 07, 2026 (0d ago) vlatest
Homepage ↗ Repo ↗ Other memory fsrs spaced-repetition knowledge-graph rust vector-search mcp local open-source sqlite
⚙ Agent Friendliness
80
/ 100
Can an agent use this?
🔒 Security
78
/ 100
Is it safe for agents?
⚡ Reliability
72
/ 100
Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality
85
Documentation
82
Error Messages
75
Auth Simplicity
80
Rate Limits
72

🔒 Security

TLS Enforcement
88
Auth Strength
78
Scope Granularity
72
Dep. Hygiene
80
Secret Handling
75

Code versioning/history MCP tool. Read access to git history. No write operations by default. Sensitive data in commit history is a common issue — audit before exposing.

⚡ Reliability

Uptime/SLA
72
Version Stability
75
Breaking Changes
70
Error Recovery
72
AF Security Reliability

Best When

You want an AI agent with rich, persistent memory that naturally decays less-important information, deduplicates overlapping knowledge, and surfaces contextually relevant memories — especially for long-running personal or project assistants.

Avoid When

You need cloud-hosted or multi-user shared memory, or your use case is simple enough that a flat context window suffices.

Use Cases

  • Giving Claude or other AI agents persistent cross-session memory with intelligent forgetting curves rather than unlimited storage
  • Building agents that remember project decisions, user preferences, and codebase patterns with deduplication and retroactive importance scoring
  • Visualizing an AI agent's knowledge graph in a real-time 3D dashboard for debugging or exploration

Not For

  • Use cases requiring cloud sync or multi-device memory sharing (fully local/offline only)
  • Lightweight deployments where the ~130MB embedding model download is prohibitive
  • Teams who need a simple key-value store rather than a cognitively sophisticated memory system

Interface

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

Authentication

OAuth: No Scopes: No

No authentication required. Fully local. Optional SQLCipher encryption for the database.

Pricing

Model: open_source
Free tier: Yes
Requires CC: No

AGPL-3.0 open source. No cloud costs — fully offline after initial embedding model download.

Agent Metadata

Pagination
cursor
Idempotent
Partial
Retry Guidance
Not documented

Known Gotchas

  • First run downloads ~130MB embedding model from Hugging Face — requires internet access
  • AGPL-3.0 license may require commercial users to open-source integrations
  • 21 MCP tools is a large surface area — agents may struggle with tool selection
  • dream consolidation runs during idle periods — agents calling this explicitly may see variable latency
  • WebSocket dashboard port separate from MCP stdio — both must be managed if using visualization

Alternatives

Full Evaluation Report

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

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 2026-03-07.

6228
Packages Evaluated
26150
Need Evaluation
173
Need Re-evaluation
Community Powered