Skip to main content
Glama
mnemostack
by udjin-labs

Mnemostack Search

mnemostack_search

Search indexed memories with hybrid recall to get raw ranked matches instead of synthesized answers; filter by payload, limit by token budget, and hide stale facts by default.

Instructions

Search indexed memories with hybrid recall.

Read-only. Requires the read scope when authentication is enabled. Use this when you need raw memory matches rather than a synthesized answer. Returns a JSON object with ok, query, count, results, tokens_estimate (estimated total text tokens of the results), notes (the AUTHORITATIVE list of routine signals for stages that did not apply — e.g. temporal:no_parse on any query without a date; NOT a fault), and degraded (components that actually fell back, plus a deprecated back-compat duplicate of the routine tags until the next major; an entry absent from notes is a real fault). Results are ranked by reciprocal rank fusion of BM25, semantic, graph, and temporal retrievers when configured; each result includes id, text, score, sources, payload, and annotations. Stale facts (invalidated_at set) are hidden by default; use include_invalidated or as_of to see them.

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": {
      "query": "<query>"
    },
    "name": "mnemostack_search"
  }
}

Input Schema

NameRequiredDescriptionDefault
as_ofNoPoint-in-time recall (ISO-8601): return facts valid at this world-time instant, ignoring later invalidation.
limitNoMaximum number of results to return (default 10)
queryYesNatural language question or keyword to search memories for
filtersNoPayload filters applied inside every retriever: exact match ({"tenant": "a"}) or gte/lte range ({"timestamp": {"gte": "2026-01-01"}}). Results never include points outside the filtered scope.
token_budgetNoHard cap on the total (estimated) text tokens of the returned results; the final ranking is cut to the prefix that fits. Unset uses the server-wide default, if any.
include_traceNoInclude the per-retriever recall trace (debug; verbose)
include_invalidatedNoInclude facts marked stale. Default false: memories with an invalidated_at marker are hidden from recall.

Output Schema

NameRequiredDescriptionDefault

No arguments

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed1 schema field changedv3.3.0
    • addedInput schema / properties / limit / minimum
      Added value: +1
  2. Changed3 schema fields changedv0.8.0
    • addedInput schema / properties / as_of
      Added value: +{
      +  "anyOf": [
      +    {
      +      "type": "string"
      +    },
      +    {
      +      "type": "null"
      +    }
      +  ],
      +  "default": null,
      +  "description": "Point-in-time recall (ISO-8601): return facts valid at this world-time instant, ignoring later invalidation."
      +}
    • addedInput schema / properties / include_invalidated
      Added value: +{
      +  "default": false,
      +  "description": "Include facts marked stale. Default false: memories with an invalidated_at marker are hidden from recall.",
      +  "type": "boolean"
      +}
    • addedInput schema / properties / token_budget
      Added value: +{
      +  "anyOf": [
      +    {
      +      "minimum": 1,
      +      "type": "integer"
      +    },
      +    {
      +      "type": "null"
      +    }
      +  ],
      +  "default": null,
      +  "description": "Hard cap on the total (estimated) text tokens of the returned results; the final ranking is cut to the prefix that fits. Unset uses the server-wide default, if any."
      +}
  3. Changed2 schema fields changedv0.6.0
    • addedInput schema / properties / filters
      Added value: +{
      +  "anyOf": [
      +    {
      +      "additionalProperties": true,
      +      "type": "object"
      +    },
      +    {
      +      "type": "null"
      +    }
      +  ],
      +  "default": null,
      +  "description": "Payload filters applied inside every retriever: exact match ({\"tenant\": \"a\"}) or gte/lte range ({\"timestamp\": {\"gte\": \"2026-01-01\"}}). Results never include points outside the filtered scope."
      +}
    • addedInput schema / properties / include_trace
      Added value: +{
      +  "default": false,
      +  "description": "Include the per-retriever recall trace (debug; verbose)",
      +  "type": "boolean"
      +}
  4. Changed2 schema fields changedv0.4.3
    • addedInput schema / properties / limit / description
      Added value: +"Maximum number of results to return (default 10)"
    • addedInput schema / properties / query / description
      Added value: +"Natural language question or keyword to search memories for"
  5. First observedv0.4.1

TDQS

A4.2/5.0
Behavior5/5

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

With no annotations supplied, the description carries the full burden and does so: it declares read-only semantics, the required OAuth scope, the default suppression of stale facts (invalidated_at), and the ranking mechanism (reciprocal rank fusion over BM25, semantic, graph, temporal retrievers when configured). It also disambiguates routine vs. real faults across notes/degraded, which prevents an agent from misreading a normal no-parse signal as a failure. That is substantive disclosure beyond anything structured fields provide.

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

Conciseness3/5

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

Front-loading is good — purpose and read-only/auth status come first. However, the notes/degraded parenthetical is a dense, hard-to-parse run-on about a deprecation back-compat duplicate that reads like release notes rather than invocation guidance, and it dilutes the more decision-relevant information. Trimming that block would improve signal without losing anything an agent needs to call the tool.

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 7-parameter search tool with an output schema and no annotations, the definition covers the essentials: safety profile, auth requirement, staleness default, ranking behavior, and result contents. It is slightly over-complete on return-value minutiae the output schema already governs, but nothing needed to invoke it correctly is missing.

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

Parameters3/5

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

Schema coverage is 100%, so every parameter including as_of, include_invalidated, filters, and token_budget is already documented in the schema with defaults and semantics. The description reinforces the stale-fact escape hatches (include_invalidated, as_of) but adds no format, interaction, or precedence detail beyond what the schema states. Baseline 3 is appropriate when the schema does the heavy lifting.

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?

Opens with a specific verb and resource ("Search indexed memories with hybrid recall") and immediately differentiates itself from the synthesized-answer sibling: "raw memory matches rather than a synthesized answer." An agent can route between retrieval and answer generation without opening any schema.

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?

The description gives a clear selection condition ("Use this when you need raw memory matches rather than a synthesized answer"), which implicitly routes to mnemostack_answer for synthesized responses. It also states the auth precondition ("Requires the read scope when authentication is enabled"). It stops short of naming the alternative tool explicitly or describing when-not to search at all.

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