Pryzm Memory

Save a memory

memory_store

Saves a memory (note, decision, preference, to-do…) in the user's account and returns its id and any close matches.

content: the text to save (max 20,000 characters).
type: technical_decision | bug | preference | intention | context | note | todo.
project: the project it belongs to; tags: optional keywords; importance: 1 to 10.
on_similar: smart (default) | new | reinforce | skip.
status: "proposal" for something suggested but not agreed, "decision" for something agreed or final.
supersedes: ids of saved memories that this one supersedes (for example the previous amount, listed in warnings); they stay visible, marked as replaced. Additive: an exact repeat is reinforced, anything else creates a new memory; close matches are returned in similar, never overwritten (supersedes only marks older memories as replaced, their text is kept). Refused by the server, with nothing saved: content starting with ## (private mode), and content containing a payment card number, an API key or private key, or a password.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "content": "<content>"
    },
    "name": "memory_store"
  }
}

Input Schema

Name	Required	Default
`tags`	No
`type`	No	note
`force`	No
`status`	No
`content`	Yes
`project`	No
`importance`	No
`on_similar`	No	smart
`supersedes`	No

Schema Changelog

Changes observed during successful MCP inspections.

First observedOct 17, 2026

TDQS

A4.5/5.0

Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Beyond the annotations it discloses additive semantics (exact repeat reinforced, similar matches returned but never overwritten), the effect of `supersedes` (older memories stay visible and only marked replaced), and explicit server-side refusals (private-mode `##` content, card numbers, API keys, passwords). This is exactly the behavioral context the readOnly/destructive/idempotent hints cannot convey.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with a one-sentence summary, then a compact parameter list and two paragraphs on additive behavior and refusals. Dense but every line carries information; only the on_similar enumeration is listed without payoff.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a non-destructive but mutating write tool with no output schema, it covers the return value, deduplication behavior, supersede semantics, and refusal conditions. The remaining gaps are minor: the unexplained `force` flag and no guidance on required permissions or account-level limits.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

With 0% schema description coverage, the description carries the full burden and documents eight of nine parameters with meaning: content size cap, the type taxonomy, project/tags/importance, the on_similar modes, status values, and supersedes semantics. The `force` parameter is left unexplained, and the on_similar modes are named but not individually defined.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description opens with a specific verb and resource ("Saves a memory") plus the content kinds it covers, and states the observable outcome ("returns its id and any close matches"). That distinguishes it immediately from read-oriented siblings like memory_get, memory_search, and memory_list.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

It gives clear context for use (saving notes, decisions, preferences, to-dos) and specifies when `status` should be "proposal" versus "decision". It stops short of explicitly naming alternatives or telling the agent when not to call it, so it is clear but not fully routing.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All Pryzm Memory tools Inspect

Try in Browser

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources

Need Help?