Nexusyn – Long-term memory engine for AI agents (Go, DiskANN, MCP)

Stop building agents that forget. Give your LLMs persistent, grounded memory across sessions.

Quickstart • MCP Integration • Architecture • Benchmarks • Cloud vs Self-Hosted • Documentation

Most AI agent frameworks implement memory by taking the last conversation turn, calculating an embedding, and storing it in a vector database.

When the agent queries this "memory", it runs a simple nearest-neighbor search. This breaks down in production:

1. Vectors don't understand time: If a user says*"I live in Berlin"* in January and*"I moved to Madrid"* in August, both vectors have identical semantic similarity to*"Where do I live?"* . A pure vector search frequently returns the outdated fact.
2. Context window bloat: Dumping raw matching chunks into the prompt forces the LLM to resolve contradictions, inflating token costs and increasing hallucination rates.
3. No entity understanding: Vector similarity misses relational facts that don't share keywords (e.g. connecting a bug report to the architecture decision that caused it).

Nexusyn is a dedicated data plane engine designed specifically for agentic long-term memory:

• 🕓 Bi-Temporal Versioning: Every fact is stamped with transaction time and valid time (valid_from /valid_to ). New statements supersede older ones automatically without deleting history.
• 🔀 Hybrid Retrieval (RRF): Fuses dense vector similarity (pgvector HNSW) + BM25 full-text search + entity graph expansion + recency & date-anchor boosts using Reciprocal Rank Fusion.
• 🎯 Reranking: High-precision cross-encoder re-scoring of candidate chunks before generation.
• 🛡️ Grounded Synthesis: Returns a synthesized, ready-to-use answer that cites verifiable source chunks — so the agent receives answers, not just raw text fragments.
• ⚡ MCP-Native: First-class Model Context Protocol server (/v1/mcp ) running over streamable HTTP. Connects directly toClaude Code ,Cursor ,Windsurf , or custom agent frameworks with zero glue code.
• 🏢 Multi-Tenant with Row-Level Security (RLS): True isolation at the PostgreSQL layer. A tenant can never see or search another tenant's memories.
• ⚙️ Async Pipeline (River): Heavy background workloads (chunking, batch embedding, entity extraction, and wiki compilation) run asynchronously on a Postgres-backed queue.

Nexusyn is evaluated against standard public benchmarks for agent long-term memory under reproducible conditions:

                    ┌─────────────────────────┐
                    │       AI Agent / IDE    │
                    │ (Claude Code, Cursor, …)│
                    └────────────┬────────────┘
                        HTTP     │   MCP (/v1/mcp)
                                 ▼
┌────────────────────────────────────────────────────────────────────────┐
│                             NEXUSYN ENGINE                             │
│                                                                        │
│   POST /v1/ingest                                POST /v1/query        │
│         │                                              │               │
│         ▼                                              ▼               │
│   [ Deduplication ]                             [ Sub-query Gen ]      │
│         │                                              │               │
│   [ Chunker ]                                   [ Hybrid Retrieval ]   │
│         │                                       ├── Vector (HNSW)      │
│         ▼                                       ├── BM25 Full-Text     │
│   [ River Queue (Async) ]                       ├── Entity Graph       │
│   ├── Batch Embeddings                          └── Date Anchor Boost  │
│   ├── Entity & Relation Extraction                     │               │
│   └── Wiki / Profile Compilation                       ▼               │
│         │                                       [ RRF Fusion ]         │
│         ▼                                              │               │
│   [ PostgreSQL 18 (pgvector) ]                  [ Cross-Encoder Rerank]│
│   Row-Level Security (Multi-Tenant)                    │               │
│                                                        ▼               │
│                                                 [ Grounded Answer ]    │
│                                                 (Synthesized + Sources)│
└────────────────────────────────────────────────────────────────────────┘

Get the complete stack running locally (PostgreSQL 17+ with pgvector, the Nexusyn HTTP/MCP API on :8044, and the River background worker) in under 60 seconds:

git clone https://github.com/nexusyn/engine.git
cd engine

cp .env.example .env


docker compose up -d

curl http://localhost:8044/health

Nexusyn runs an HTTP Model Context Protocol (MCP) server at /v1/mcp. Any MCP-compliant client gets persistent long-term memory tools out of the box.

Connect Nexusyn to Claude Code with a single CLI command:

claude mcp add nexusyn --transport http \
  "http://localhost:8044/v1/mcp" \
  --header "Authorization: Bearer <YOUR_API_TOKEN>"

Add Nexusyn to your Cursor global or project configuration:

{
  "mcpServers": {
    "nexusyn": {
      "url": "http://localhost:8044/v1/mcp",
      "headers": {
        "Authorization": "Bearer <YOUR_API_TOKEN>"
      }
    }
  }
}
{
  "mcpServers": {
    "nexusyn": {
      "serverUrl": "http://localhost:8044/v1/mcp",
      "headers": {
        "Authorization": "Bearer <YOUR_API_TOKEN>"
      }
    }
  }
}

• add_memory — Record a decision, convention, lesson, or fact tagged byproject andagent .
• search_memory — Query memories using hybrid search, returning grounded answers with sources.
• get_guideline — Retrieve organization/project-wide mandatory standards.
• update_memory /delete_memory — In-place curation and correction of existing records.

Two endpoints cover 90% of all integration needs:

curl -X POST http://localhost:8044/v1/ingest \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "title": "Auth Architecture",
    "content": "We migrated from session cookies to stateless JWT with Ed25519 token signing on 2026-08-15.",
    "agent": "cursor",
    "project": "mobile-app"
  }'

Response:

{
  "job_id": 482,
  "status": "queued"
}
curl -X POST http://localhost:8044/v1/query \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "question": "What token signing algorithm do we use for authentication?",
    "project": "mobile-app"
  }'

Response:

{
  "question": "What token signing algorithm do we use for authentication?",
  "answer": "The authentication system uses stateless JWTs signed with the Ed25519 algorithm, migrated on August 15, 2026.",
  "sources": [
    {
      "id": "chk_9a8b7c",
      "title": "Auth Architecture",
      "content": "We migrated from session cookies to stateless JWT with Ed25519 token signing on 2026-08-15.",
      "score": 0.942
    }
  ],
  "usage": {
    "model": "gpt-4o-mini",
    "prompt_tokens": 312,
    "completion_tokens": 38
  }
}

Streaming: Real-time token streaming with Server-Sent Events (SSE) is available at POST /v1/query/stream.

Nexusyn uses an adapter architecture. Bring your own models for generation, embeddings, and reranking:

go build -v ./...

go test -v ./internal/core/... ./internal/auth/... ./internal/dateutil/...

gitleaks detect --source . --no-git
govulncheck ./...

• Bug Reports & Feature Requests: Please open an issue onGitHub Issues .

• Security Vulnerabilities: Review ourSecurity Policy and report privately to**security@nexusyn.ai** .

• Website:https://nexusyn.ai

• Documentation:https://nexusyn.ai/docs

• Source Code: Released under theApache License, Version 2.0 .

• Trademark: "Nexusyn" and associated marks are trademarks of RedFoxCode.