# Synapse Layer

> AI Agent Memory Infrastructure — Trust Layer for the Agentic Internet
> Version: 2.4.6
> Full spec: https://synapselayer.org/llms-full.txt

Synapse Layer is persistent, encrypted memory for AI agents. Anything an agent
learns — decisions, preferences, project rules — is stored, protected, and
shared between agents across apps, models, and sessions. It is a trust layer
for agents, not a database for humans.

## When to use

- Give agents memory that survives across sessions, models, and tools
- Hand off context between agents (store → recall) with full audit trail
- Enforce security by default: encryption at rest, PII sanitization, intent validation
- Comply with LGPD/GDPR while sharing memory across your own agents

## Core Capabilities

- **Persistent Memory**: Store and recall context across agent sessions
- **AES-256-GCM Encryption**: Server-side encryption at rest with per-operation random IV
- **MCP Native Protocol**: First-class Model Context Protocol integration
- **Neural Handover**: One agent stores, another recalls — with full audit trail
- **Trust Quotient**: Per-memory confidence scoring (0.0–1.0)
- **Semantic Recall**: pgvector HNSW-powered similarity search
- **Immutable Audit Trail**: Three-layer lifecycle (live data, tombstone, analytics)
- **OAuth 2.0 + PKCE S256**: Industry-standard authorization
- **LGPD / GDPR Aligned**: Hard-delete on request, consent-gated storage

## Quick Start

```
pip install synapse-layer
```

```python
from synapse_layer import Synapse
client = Synapse(token="sk_connect_...")
client.store(content="User prefers dark mode", memory_type="long_term")
memories = client.recall(query="user theme preference", top_k=5)
recent = client.list_memories(limit=10)   # metadata + governance pagination
```

Full SDK API reference (Python `Synapse`/`SynapseClient` + TypeScript
`SynapseClient`): https://synapselayer.org/llms-full.txt#sdk-api-reference · https://synapselayer.org/docs#sdk-api

## Supported Platforms

- Claude Desktop — MCP native
- Claude API — MCP / REST
- Cursor — MCP integration
- LangChain — Python SDK
- Vercel AI SDK — TypeScript SDK
- CrewAI, AutoGen, LangGraph — Coming Soon
- n8n, Zapier — Coming Soon

## Interop & Registries

- [MCP Marketplace](https://mcp-marketplace.io/server/io-github-synapselayer-synapse-layer): Official MCP server listing
- [Smithery](https://smithery.ai/server/synapselayer/synapselayer): MCP server registry listing
- [PyPI](https://pypi.org/project/synapse-layer/): Python SDK package (v2.4.6)
- [npm](https://www.npmjs.com/package/synapse-layer): TypeScript SDK package
- [GitHub](https://github.com/SynapseLayer/synapse-layer): Source code, issues, and contributions

## Pricing

- Free: $0/month — AES-256-GCM, REST API, MCP protocol
- Pro: $19/month — Neural Handover, OAuth + PKCE, SDK access, email support
- Enterprise: Contact sales — custom deployment, dedicated support

## Links

- [Full Specification](https://synapselayer.org/llms-full.txt): Complete technical spec with all capabilities, security details, and API reference.
- [Documentation](https://synapselayer.org/docs): Installation guides, API reference, and MCP configuration.
- [Connect Guide](https://synapselayer.org/connect): Self-service setup in under 5 minutes — Claude Desktop, Cursor, Codex, Hermes, DeepAgent, LangChain, n8n, SDKs.
- [OpenAPI Spec](https://synapselayer.org/api/openapi.json): Machine-readable API specification (OpenAPI 3.1).
- [Interactive API Reference](https://synapselayer.org/api-docs): Try every endpoint in the browser (Swagger UI).
- [MCP Manifest](https://synapselayer.org/.well-known/mcp.json): MCP server discovery manifest with auth, endpoint, and capabilities.
- [MCP Server Card](https://synapselayer.org/.well-known/mcp/server-card.json): Detailed MCP server capabilities, tools, and parameters.
- [Trust Manifest](https://synapselayer.org/.well-known/trust.json): Security, compliance, encryption, and discovery metadata.
- [Pricing Manifest](https://synapselayer.org/.well-known/pricing.json): Machine-readable pricing for agent commerce.
- [Agent Skill](https://synapselayer.org/skill.md): When and how to use Synapse Layer — agent-readable skill file.
- [Forge Dashboard](https://forge.synapselayer.org): Operational dashboard for memory management.
- [MCP Marketplace](https://mcp-marketplace.io/server/io-github-synapselayer-synapse-layer): MCP server listing.
- [Smithery](https://smithery.ai/server/synapselayer/synapselayer): MCP server registry listing.