Retrieve focused context
Request the record a task needs and fetch history when the question calls for it.
MCP integration
Put memory within your agent’s reach.
Give a compatible MCP client tools to record observations, retrieve current facts, and explore relationships. Connect a workspace through Streamable HTTP.
In practice
Read. Act. Remember what changed.
Illustrative tool sequence. The client decides which tools to call; workspace policy governs incoming observations.
- Before a task
- memory_get loads the relevant entity.
- When context is missing
- memory_search or memory_traverse finds related information.
- After new information
- memory_insert submits an observation.
- When a change needs explanation
- memory_history retrieves the entity changelog.
Built into the workflow
Useful tools. Governed writes.
Use your workspace rules
MCP observations pass through consolidation with ontology validation and provenance.
Keep people in the loop
Inspect changes in the console and review proposals held by policy.
Tool reference
Ten tools for your memory workflow.
memory_insertSubmit an observation (proposal).
memory_searchText search over current state.
memory_getLoad an entity’s current view.
memory_historyChangelog for an entity.
memory_traverseWalk typed relations.
ontology_defineAdd class, predicate, or relation type.
ontology_listRead the schema that gates writes.
memory_exportProjection of current state — not a write surface.
memory_retractRetract a current fact.
memory_reviewApprove or reject a held proposal.
Connect your workspace.
Create an access key in Settings, then use the Connections page for the endpoint and client configuration. Keep the key private.
Create your workspaceCOMMON QUESTIONS
MCP, answered.
What is the ConceptMem MCP server?
A Model Context Protocol server, served over Streamable HTTP, that exposes a workspace as tools a compatible client can call: memory_insert, memory_search, memory_get, memory_history, memory_traverse, ontology_define, ontology_list, memory_export, memory_retract, and memory_review.
Which clients can use it?
Any MCP client that supports Streamable HTTP. The endpoint and headers are shown on the Connections page of your workspace once you create an access key.
Are MCP writes governed the same way?
Yes. memory_insert submits an observation that passes through consolidation with ontology validation and provenance, exactly like a write from the console or the SDK.
How should an agent use the tools in a task?
Read before acting and write after learning: load the relevant entity, search or traverse when context is missing, submit an observation once new information arrives, and read the changelog when a change needs an explanation.
Does the agent need to know the schema first?
It helps. ontology_list reports the classes, predicates, relation types, and constraints that gate writes, including inherited ones, so the agent can reuse the existing vocabulary instead of inventing a parallel one.
Make your next update count
Start with one useful memory.
Create a workspace, define a type, and see your first observation become a traceable record.