Save one atomic fact to long-term memory. Server stores multiple dimensions (category, observation_type, concepts, tags) for later filter/recall. RULES: (1) ONE fact per call. (2) DATES: Include date (YYYY-MM-DD) only for event-like facts; omit for timeless facts. (3) CONTEXT: Rewrite 'it/this/th...
AI agents use save_memory to create or update resources in Memoryx — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Memoryx environment.
| Parameter | Type | Required | Description |
|---|---|---|---|
content | string | Yes | REQUIRED. One atomic fact, self-contained. Include date (YYYY-MM-DD) only for event-like facts; omit for timeless facts (preferences, definitions). MUST resolve |
metadata | object | — | Multi-dimensional metadata (server stores for filter/recall). category (required): semantic|episodic|procedural|emotional|reflective. observation_type (optional |
Parameters from the server's own tool schema.
This tool creates new records in a memory store and modifies the state of stored information. It is reversible (paired with delete_memory on the sibling tools list), making it Write rather than Destructive.
From the tool's definition save_memory creates and stores data in long-term memory with configurable metadata (category, observation_type, concepts, tags).
Risk signalsAccepts raw HTML/template content (content)
Attacks that exploit this kind of access
Save one atomic fact to long-term memory. Server stores multiple dimensions (category, observation_type, concepts, tags) for later filter/recall. RULES: (1) ONE fact per call. (2) DATES: Include date (YYYY-MM-DD) only for event-like facts; omit for timeless facts. (3) CONTEXT: Rewrite 'it/this/that' to explicit referents. (4) metadata.category (required): semantic|episodic|procedural|emotional|reflective. (5) metadata.observation_type (optional, code): bugfix|feature|refactor|change|discovery|decision. (6) metadata.concepts (optional): array of how-it-works|why-it-exists|what-changed|problem-solution|gotcha|pattern|trade-off—do not put observation_type inside concepts. (7) metadata.tags (optional): string array, e.g. conv:<conversation_id> for session scoping. (8) Call when: 'remember this', design/code decisions, task completion, end-of-session summary (≤100 words). (9) Include concrete anchors: class, method, file path, SDK, config key, error type. It is categorised as a Write tool in the Memoryx MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
save_memory accepts 2 parameters: content, metadata. Required: content. The full parameter table on this page comes from the server's own tool schema.
Register the Memoryx MCP server in PolicyLayer and add a rule for save_memory: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Memoryx. Nothing to install.
save_memory is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the save_memory rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for save_memory. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
save_memory is provided by the Memoryx MCP server (@t0ken.ai/memoryx-mcp-server). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →