synapselayer avatar

Synapse Layer — Trust Infrastructure for AI Agents

synapselayer/synapse-layer

Persistent memory for Claude and other MCP clients, backed by encrypted storage and semantic search. Exposes 13 tools including recall, save_to_synapse, and neural_handover for cross-session context. Memories are encrypted at rest with AES-256-GCM and indexed via pgvector for semantic retrieval. Works standalone via Python SDK or connects to their Forge cloud API for cross-agent memory sharing. Integrates with Claude Desktop, Cursor, LangChain, and CrewAI. Reach for this when you need your agent to remember user preferences, prior decisions, or conversation context across sessions without rebuilding RAG infrastructure from scratch.

GitHub stars
13
tools
13
auth
Secret needed
transport
HTTP
last updated
Sep 16, 2026

Install

From registry metadata
$ npx add-mcp https://forge.synapselayer.org/api/mcp --name synapse-layer --header 'x-connect-token: ${X_CONNECT_TOKEN}'

Run in your terminal. add-mcp asks which agents to add it to, then asks for each value it needs. Add -g to install it for every project.

Review the command, arguments and environment values before installing. MCP servers run with your local permissions.

Other ways

Claude Code

$ claude mcp add --transport http synapse-layer https://forge.synapselayer.org/api/mcp --header 'x-connect-token: YOUR_X_CONNECT_TOKEN'

Run in your terminal. Add --scope user to make it available in every project.

Tools

Verified live · 13 tools

Verified live against the running server on Nov 6, 2026.

recallRetrieve relevant persisted memory using semantic, temporal, priority, or hybrid routing. Governance: requires reason (10–200 chars). Rate limit: 20/min.8 params

Retrieve relevant persisted memory using semantic, temporal, priority, or hybrid routing. Governance: requires reason (10–200 chars). Rate limit: 20/min.

Parameters* required
agent_idstring
Agent identifier (canonical or alias). Used only when scope="agent" or scope is inferred.
limitnumber
Maximum memories to return (1–50, default: 10).
modestring
Recall routing mode. "temporal" = pure chronological, "semantic" = FTS + TQ ranking, "priority" = critical/high-TQ first, "hybrid" = weighted blend, "auto" = detect from query (default).one of auto · temporal · semantic · priority · hybrid
project_idstring
Optional project ID. Required for Corporate context unless the token is project-bound.
query*string
What to recall — natural language query for memory retrieval.
reason*string
REQUIRED (10–200 chars). Audit policy: MEMORY_EXPORT_GOVERNANCE_V1. Human-readable justification for retrieving PLAINTEXT memory. Without this parameter the call WILL fail with code -32602. Example...
scopestring
Explicit memory scope. "agent" (default, fail-closed) = limits recall to the given agent_id (or the token agent). "tenant" = cross-agent recall within the authenticated tenant. If omitted with agen...one of agent · tenant
subjectstring
ESCOPO1 (optional): opaque 8-128 char token ([A-Za-z0-9._:-]) that isolates memory to a single person within a shared connect-token. Omit for token-scoped behavior. Never include PII/email.
save_to_synapsePersist memory with encryption at rest, sanitization, and deduplication controls.11 params

Persist memory with encryption at rest, sanitization, and deduplication controls.

Parameters* required
agent_idstring
Agent identifier for memory isolation. Defaults to "default".
content*string
The memory content to store securely.
force_intentstring
MANUAL OVERRIDE ONLY — use apenas quando você (humano) quer forçar explicitamente a classificação de intenção. Valores reconhecidos: preference | fact | procedural | bio | critical | operational |...
importancenumber
Accepted for compatibility; not enforced by the API.
intent_reasonstring
Auditable reason for forcing intent classification. Max 500 characters. Ignored if force_intent is absent.
projectstring
Project identifier (e.g., SYNAPSE_LAYER).
project_idstring
Project ID for authorization and memory isolation. Required for Corporate context unless the token is project-bound.
subjectstring
ESCOPO1 (optional): opaque 8-128 char token ([A-Za-z0-9._:-]) that isolates memory to a single person within a shared connect-token. Omit for token-scoped behavior. Never include PII/email.
tagsarray
Tags for categorization.
typestring
Event type: [MILESTONE], [DECISION], [ALERT], [AUTO-STRAT], [AUTO-OP], [AUTO-INSIGHT], [AUTO-DECISION], [AUTO-CONTEXT], [MANUAL].
visibilitystring
private (default) or project (visible to active members of this project).one of private · project
process_textExtract candidate memories from free-form text with governance filters and sanitization.6 params

Extract candidate memories from free-form text with governance filters and sanitization.

Parameters* required
agent_idstring
Agent identifier. Defaults to "default".
projectstring
Force a specific project namespace. Auto-detected if omitted.
project_idstring
Project ID for authorization and memory isolation. Required for Corporate context unless the token is project-bound.
sourcestring
Source identifier (default: mcp).
text*string
Free-form text to scan for auto-save triggers.
visibilitystring
private (default) or project (visible to active members of this project).one of private · project
searchSearch persisted memory across agent scopes using full-text matching. Governance: requires reason (10–200 chars) because results contain plaintext content.7 params

Search persisted memory across agent scopes using full-text matching. Governance: requires reason (10–200 chars) because results contain plaintext content.

Parameters* required
agent_idstring
Agent identifier (canonical or alias). Used only when scope="agent" or scope is inferred.
limitnumber
Maximum results to return (1–50, default: 20).
project_idstring
Project ID for authorization and memory isolation. Required for Corporate context unless the token is project-bound.
query*string
Search query — natural language or keywords.
reason*string
REQUIRED (10–200 chars). Audit policy: MEMORY_EXPORT_GOVERNANCE_V1. Human-readable justification for retrieving PLAINTEXT memory. Without this parameter the call WILL fail with code -32602. Example...
scopestring
Explicit search scope. "agent" (default, fail-closed) = limits search to the given agent_id (or the token agent). "tenant" = cross-agent search within the authenticated tenant. If omitted with agen...one of agent · tenant
subjectstring
ESCOPO1 (optional): opaque 8-128 char token ([A-Za-z0-9._:-]) that isolates memory to a single person within a shared connect-token. Omit for token-scoped behavior. Never include PII/email.
health_checkCheck service availability, engine version, and storage health. Public method.

Check service availability, engine version, and storage health. Public method.

No parameters. Call it with no arguments.

initialize_contextInitialize a persistent memory context for a conversation or agent session. Returns REDACTED metadata by default (detail="summary"). Use detail="full" with a reason (10–200 chars) to retrieve plaintext content under MEMORY_EXPORT_GOVERNANCE_V1.3 params

Initialize a persistent memory context for a conversation or agent session. Returns REDACTED metadata by default (detail="summary"). Use detail="full" with a reason (10–200 chars) to retrieve plaintext content under MEMORY_EXPORT_GOVERNANCE_V1.

Parameters* required
detailstring
Detail level. "summary" (default) = redacted metadata only, no plaintext. "full" = plaintext content; REQUIRES the reason parameter and is audit-logged.one of summary · full
project_idstring
Optional project ID. Required for Corporate context unless the token is project-bound.
reasonstring
REQUIRED only when detail="full" (10–200 chars). Audit policy: MEMORY_EXPORT_GOVERNANCE_V1. Human-readable justification for retrieving PLAINTEXT context. Without it, detail="full" WILL fail with c...
save_memorySave a memory entry to the persistent store. Alias of save_to_synapse.8 params

Save a memory entry to the persistent store. Alias of save_to_synapse.

Parameters* required
agent_idstring
Agent identifier for memory isolation.
content*string
The memory content to store securely.
force_intentstring
MANUAL OVERRIDE ONLY — use apenas quando você (humano) quer forçar explicitamente a classificação de intenção. Valores reconhecidos: preference | fact | procedural | bio | critical | operational |...
intent_reasonstring
Auditable reason for forcing intent classification. Max 500 characters. Ignored if force_intent is absent.
project_idstring
Project ID for authorization and memory isolation. Required for Corporate context unless the token is project-bound.
subjectstring
ESCOPO1 (optional): opaque 8-128 char token ([A-Za-z0-9._:-]) that isolates memory to a single person within a shared connect-token. Omit for token-scoped behavior. Never include PII/email.
tagsarray
Tags for categorization.
visibilitystring
private (default) or project (visible to active members of this project).one of private · project
store_memoryStore structured memory with metadata and trust scoring.8 params

Store structured memory with metadata and trust scoring.

Parameters* required
agent_idstring
Agent identifier for memory isolation. Defaults to "default".
content*string
The memory content to store securely.
force_intentstring
MANUAL OVERRIDE ONLY — use apenas quando você (humano) quer forçar explicitamente a classificação de intenção. Valores reconhecidos: preference | fact | procedural | bio | critical | operational |...
intent_reasonstring
Auditable reason for forcing intent classification. Max 500 characters. Ignored if force_intent is absent.
project_idstring
Project ID for authorization and memory isolation. Required for Corporate context unless the token is project-bound.
subjectstring
ESCOPO1 (optional): opaque 8-128 char token ([A-Za-z0-9._:-]) that isolates memory to a single person within a shared connect-token. Omit for token-scoped behavior. Never include PII/email.
tagsarray
Tags for categorization.
visibilitystring
private (default) or project (visible to active members of this project).one of private · project
recall_memoryRecall persisted memory by query. Alias of recall.8 params

Recall persisted memory by query. Alias of recall.

Parameters* required
agent_idstring
Agent identifier (canonical or alias). Used only when scope="agent" or scope is inferred.
limitnumber
Maximum memories to return (1–50, default: 10).
modestring
Recall routing mode.one of auto · temporal · semantic · priority · hybrid
project_idstring
Optional project ID. Required for Corporate context unless the token is project-bound.
query*string
What to recall — natural language query for memory retrieval.
reason*string
REQUIRED (10–200 chars). Audit policy: MEMORY_EXPORT_GOVERNANCE_V1. Without this the call WILL fail with code -32602. Example: "Checking user project status."
scopestring
Explicit memory scope. "agent" (default, fail-closed) = limits recall to the given agent_id (or the token agent). "tenant" = cross-agent recall within the authenticated tenant. If omitted with agen...one of agent · tenant
subjectstring
ESCOPO1 (optional): opaque 8-128 char token ([A-Za-z0-9._:-]) that isolates memory to a single person within a shared connect-token. Omit for token-scoped behavior. Never include PII/email.
list_memoriesList memory metadata with pagination and governance limits.5 params

List memory metadata with pagination and governance limits.

Parameters* required
agent_idstring
Agent identifier (canonical or alias). Used only when scope="agent" or scope is inferred.
limitnumber
Maximum memory metadata rows to return. Hard-capped: default 5, max 10. Values above 10 are silently reduced to 10.
project_idstring
Optional project ID. Required for Corporate context unless the token is project-bound.
scopestring
Explicit memory scope. "agent" (default, fail-closed) = lists memories for the given agent_id (or the token agent). "tenant" = lists memories across all agents of the authenticated tenant.one of agent · tenant
subjectstring
ESCOPO1 (optional): opaque 8-128 char token ([A-Za-z0-9._:-]) that isolates memory to a single person within a shared connect-token. Omit for token-scoped behavior. Never include PII/email.
memory_feedbackSubmit feedback to adjust trust scoring for stored memories.6 params

Submit feedback to adjust trust scoring for stored memories.

Parameters* required
memoryIdstring
Deprecated camelCase alias of memory_id — still accepted for backward compatibility.
memory_idstring
ID of the memory to provide feedback on (canonical snake_case name).
project_idstring
Project ID for authorization and memory isolation. Required for Corporate context unless the token is project-bound.
sessionIdstring
Deprecated camelCase alias of session_id — still accepted for backward compatibility.
session_idstring
Optional session identifier for tracking (canonical snake_case name).
signal*string
Feedback signal: used (auto), helpful (positive), irrelevant (negative), ignored.one of used · ignored · helpful · irrelevant
neural_handoverTransfer contextual state between agents with continuity controls.4 params

Transfer contextual state between agents with continuity controls.

Parameters* required
consuming_agentstring
Agent identifier consuming the handover (defaults to token agent).
project_idstring
Project ID for authorization and memory isolation. Required for Corporate context unless the token is project-bound.
reason*string
REQUIRED (10–200 chars). Human-readable justification for consuming this handover. Example: "Receiving project context from Claude session."
token*string
The handover token (64 hex chars). Generated by the source agent/user in Forge UI.
slo_reportReturn uptime and SLO metrics for the MCP service.2 params

Return uptime and SLO metrics for the MCP service.

Parameters* required
admin_token*string
Admin authentication token. Must match ADMIN_TOKEN environment variable.
window_hoursarray
Time windows in hours for the report. Default: [24, 168] (24h + 7d).

Sponsors

Readme

🧠 Synapse Layer

RAG retrieves. Synapse remembers.

Persistent memory infrastructure for AI agents — AES-256-GCM encrypted at rest, semantic search, MCP-native.

Synapse Layer is open-source persistent memory infrastructure for AI agents and assistants. Memories are encrypted at rest with AES-256-GCM, indexed via pgvector HNSW for semantic recall, and exposed through MCP JSON-RPC for native integration with Claude, GPT, Gemini, and any MCP-compatible client. Apache 2.0 licensed.

Status and limits: see FEATURES.md — the public contract for what is shipped, gated, experimental or roadmap.


View the full README on GitHub

Details

Repository
github.com/synapselayer/synapse-layer
Website
synapselayer.org
Categories
AI & LLM Tools, Cloud & Infrastructure
Transport
HTTP
Auth
Required
Registry
MCP registry, active
Tools verified
Nov 6, 2026

Related AI & LLM Tools MCP servers

View all
1hashgraph-online avatarHashnethashgraph-online/hashnet-mcp-jsUniversal AI Agent Discovery - search 59,000+ agents across MCP, NANDA, A2A, OpenRouter0stars
2securityronin avatarAlaya Mcpsecurityronin/alayaLocal memory engine for AI agents with knowledge graphs, forgetting, and semantic recall12stars
3meloncafe avatarChromaDB Remote MCP Servermeloncafe/chromadb-remote-mcpRemote ChromaDB vector database MCP server with streamable HTTP transport12stars
4var-gg avatarPindocvar-gg/pindocCode-pinned team memory for AI coding agents — typed artifacts, MCP-native, self-host.12stars
5haseebkhalid1507 avatarVelociraghaseebkhalid1507/velociragLightning-fast RAG for AI agents. 4-layer fusion, ONNX Runtime, sub-200ms search.10stars
6octen-team avatarOcten Mcpocten-team/octen-mcpURL to LLM-ready markdown — plus per-page category, page_structure, and query-driven highlights.10stars