Documentation

Everything you need to build memory-enabled agents with Agentverse Memory.

Getting Started, MCP Integration, and full 35-tool API Reference are live
🚀

Getting Started

live

Connect your first agent to Agentverse Memory in under 5 minutes.

🔌

MCP Integration

live

35 tools over JSON-RPC 2.0. Works with Claude Desktop, Cursor, LangChain, CrewAI.

🐍

Python SDK

soon

pip install agentverse-memory — full async support, type hints, 96% entity recall.

📘

TypeScript SDK

soon

npm install @fetchai/agentverse-memory — ESM + CJS, full TypeScript types.

📚

API Reference

live

35 MCP tools documented with parameters, return types, and examples.

🏛️

Architecture

soon

How GraphPalace works: TF-IDF extraction, LadybugDB graph engine, pheromone retrieval.

5-minute quickstart (Python)

Full guide →
# 1. Install
pip install agentverse-memory

# 2. Connect
from agentverse_memory import MemoryClient

client = MemoryClient(api_key="your-api-key")
agent = client.agent("my-agent-id")

# 3. Write — no LLM calls; sub-ms in-process engine write op
agent.add("Alice approved the Q3 budget on Tuesday")
agent.add("Bob manages Alice and owns account #204")

# 4. Recall — semantic search
memories = agent.recall("budget decisions")

# 5. Graph query — multi-hop reasoning
result = agent.graph_query(
    "MATCH (a:Person)-[:MANAGES]->(b:Person) RETURN a, b"
)

# 6. Pheromone retrieval — frequency-biased
hot = agent.recall("Q3", mode="pheromone")

API Reference 35 MCP tools

EpisodicSemanticGraphProceduralWorkingPheromoneSpacesSystemTraversal
Episodicmemory_store_episodeStore a time-stamped event. No LLM at write time — sub-ms in-process engine write op.
Episodicmemory_get_episodesRetrieve episodes with filtering and pagination
Episodicmemory_search_episodesNatural-language search over episodes with pheromone weighting
Episodicmemory_search_timelineTime-windowed episode search
Episodicmemory_consolidate_episodesMerge related episodes into a consolidated record
Semanticmemory_store_entityStore a named entity with type and metadata
Semanticmemory_get_entityRetrieve a specific entity by name or ID
Semanticmemory_list_entitiesList all entities with optional type/metadata filters
Semanticmemory_store_relationStore a directed relationship between entities
Semanticmemory_get_relationsGet all relations for a given entity
Graphmemory_query_graphStructured query over the knowledge graph
Graphmemory_semantic_searchSemantic similarity search across all memory types
Graphmemory_get_neighborsGet immediate neighbors of a node
Graphmemory_graph_add_tripleAdd a (subject, predicate, object) triple directly
Graphmemory_graph_neighborsExtended neighbor query with depth and filters
Graphmemory_graph_shortest_pathFind shortest path between two entities
Proceduralmemory_store_procedureStore a skill definition: goal → steps → outcome
Proceduralmemory_get_procedureRetrieve a procedure by ID
Proceduralmemory_match_procedureSearch procedures by goal description
Proceduralmemory_update_procedureUpdate steps or log execution outcome
Workingmemory_set_workingSet a key-value pair with optional TTL
Workingmemory_get_workingGet a working memory value
Workingmemory_list_workingList all active working memory keys
Workingmemory_clear_workingClear all working memory
Pheromonememory_deposit_pheromoneManually deposit a pheromone trace on a node or edge
Pheromonememory_get_pheromoneRead current pheromone weight (decay-adjusted)
Spacesmemory_create_shared_spaceCreate a shared memory space with JWT/DID access control
Spacesmemory_join_shared_spaceJoin an existing space via invite token
Spacesmemory_shared_store_entityStore an entity into a shared space
Spacesmemory_shared_queryQuery across all agents in a space
Spacesmemory_list_shared_spacesList spaces the agent belongs to
Systemmemory_get_statsGet usage stats, op counts, rate limit status
Systemmemory_delete_agentPermanently delete all memories for an agent
Traversalmemory_find_pathA* pathfinding between two concepts (all tiers)
Traversalmemory_traverse_graphMulti-hop pheromone-weighted graph traversal

Full parameter documentation, request/response schemas, and code examples coming soon. Join Discord for early access.

Questions? We're in Discord.

The team responds daily. Ask about integration, architecture, or roadmap.

Join Discord