total-agent-memory

Overview Schema Source Related Servers Discussions Admin

memory_recall

Read-onlyIdempotent

Instructions

Search ALL memory: decisions, solutions, facts, lessons from ALL past sessions. 6-stage pipeline: FTS5+BM25 → semantic → fuzzy → graph → (optional) CrossEncoder → (optional) MMR. Default: hybrid mode (BM25 + semantic + RRF). Use BEFORE starting any task. v11.0: routes to fast hot path when MEMORY_MODE=fast (default). Use memory_search_fast / memory_explain_search for explicit fast routing.

Input Schema

Name	Required	Description	Default
`mode`	No	Progressive-disclosure mode: 'search' (default) = normal results, 'index' = ultra-compact metadata only (id+title+score+type+project+created_at, ~40-60 tok/hit, no cognitive expansion, use memory_get(ids=...) to fetch full content), 'timeline' = chronological compact view; 'context' = source excerpts; 'evidence' = indexed passage search with bounded follow-up retrieval	search
`type`	No		all
`as_of`	No	ISO date or datetime that relative dates in the query ('last month', 'three weeks ago', 'yesterday') are counted from. Default: now.
`limit`	No
`query`	Yes	What to search for
`branch`	No	Filter by git branch (also includes branch-agnostic records)
`detail`	No	Level of detail: 'compact' ~50 tokens/result (id+title+score), 'summary' truncates content to 150 chars, 'full' returns everything, 'auto' picks based on query complexity (paths/urls/code → full, short → compact). Ignored when mode!='search'.	full
`fusion`	No	Score fusion method: 'rrf' = Reciprocal Rank Fusion (better multi-tier ranking), 'legacy' = original additive scoring	rrf
`intent`	No	Filter by classified intent (question\|procedural\|fact\|decision\|problem\|solution\|incident\|plan)
`rerank`	No	Enable CrossEncoder re-ranking for higher precision (adds ~30ms latency)
`topics`	No	Filter results to records tagged with any of these topics (from deep enrichment)
`diverse`	No	Enable MMR diversity to reduce redundant results (useful for broad queries)
`project`	No	Filter by project name
`entities`	No	Filter by extracted entity names (technology/person/project, case-insensitive)
`neighbors`	No	Timeline/context modes: records before/after each hit; context accepts 0–3.
`fill_budget`	No	Context mode: search deeper (up to 100 hits) and keep whole hits in rank order until context_max_chars is used, instead of excerpting the top `limit` hits to fit. Neighbours default to 0 here; pass `neighbors` to keep each hit with the records around it.
`expand_budget`	No	Max number of additional records to include via graph expansion
`decisions_only`	No	Return only structured decisions (v8.0): type=decision AND tags contain 'structured'. Results include parsed schema payload under 'decision'.
`expand_context`	No	Add graph-related records (1-hop neighbors via knowledge graph) as 'expansion' results
`missing_relation`	No	Evidence mode: missing relation; subject must occur in the original question.
`context_max_chars`	No	Context character budget; evidence mode uses the same value as a stricter UTF-8 byte budget.
`evidence_followup`	No	Evidence mode: allow one search for an explicitly supplied missing_relation.

Schema Changelog

Changes observed during successful MCP inspections.

Changed1 schema field changedv14.9.0Oct 15, 2026

addedInput schema / properties / as_of

Added value: +{
+  "description": "ISO date or datetime that relative dates in the query ('last month', 'three weeks ago', 'yesterday') are counted from. Default: now.",
+  "type": "string"
+}

Changed1 schema field changedv14.7.0Sep 30, 2026

addedInput schema / properties / fill_budget

Added value: +{
+  "default": false,
+  "description": "Context mode: search deeper (up to 100 hits) and keep whole hits in rank order until context_max_chars is used, instead of excerpting the top `limit` hits to fit. Neighbours default to 0 here; pass `neighbors` to keep each hit with the records around it.",
+  "type": "boolean"
+}

Changed6 schema fields changedv14.0.0Sep 17, 2026

addedInput schema / properties / context_max_chars

Added value: +{
+  "default": 24000,
+  "description": "Context character budget; evidence mode uses the same value as a stricter UTF-8 byte budget.",
+  "minimum": 512,
+  "type": "integer"
+}

addedInput schema / properties / evidence_followup

Added value: +{
+  "default": true,
+  "description": "Evidence mode: allow one search for an explicitly supplied missing_relation.",
+  "type": "boolean"
+}

addedInput schema / properties / missing_relation

Added value: +{
+  "additionalProperties": false,
+  "description": "Evidence mode: missing relation; subject must occur in the original question.",
+  "properties": {
+    "relation": {
+      "type": "string"
+    },
+    "subject": {
+      "type": "string"
+    },
+    "time": {
+      "type": "string"
+    }
+  },
+  "required": [
+    "subject",
+    "relation"
+  ],
+  "type": "object"
+}

changedInput schema / properties / mode / description

Previous value: -"Progressive-disclosure mode: 'search' (default) = normal results, 'index' = ultra-compact metadata only (id+title+score+type+project+created_at, ~40-60 tok/hit, no cognitive expansion, use memory_get(ids=...) to fetch full content), 'timeline' = top-K hits expanded with ±neighbors from same session (chronological)"New value: +"Progressive-disclosure mode: 'search' (default) = normal results, 'index' = ultra-compact metadata only (id+title+score+type+project+created_at, ~40-60 tok/hit, no cognitive expansion, use memory_get(ids=...) to fetch full content), 'timeline' = chronological compact view; 'context' = source excerpts; 'evidence' = indexed passage search with bounded follow-up retrieval"

changedInput schema / properties / mode / enum

Previous value: -[
-  "search",
-  "index",
-  "timeline"
-]New value: +[
+  "search",
+  "index",
+  "timeline",
+  "context",
+  "evidence"
+]

changedInput schema / properties / neighbors / description

Previous value: -"Timeline mode only: how many records before/after each hit to include."New value: +"Timeline/context modes: records before/after each hit; context accepts 0–3."

First observedv0.1.0Sep 7, 2026

TDQS

A4.1/5.0

Behavior4/5

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

Annotations already declare readOnly/idempotent/non-destructive/openWorld=false, so the safety profile is covered. The description adds genuinely useful behavior beyond that: the 6-stage pipeline order, the default hybrid mode, latency of CrossEncoder rerank, and MEMORY_MODE=fast hot-path routing. Return format is not described, keeping it from a 5.

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-loads the core capability and scope, then layers pipeline and routing details. Mostly every sentence earns its place, though the version-tagged pipeline enumeration is dense and borders on implementation trivia for a selection decision.

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 22-parameter tool with 91% schema coverage and no output schema, the description supplies enough high-level context to call it correctly: scope, default mode, and routing behavior. It omits return-shape and mode-selection guidance at the level some agents might want, but the rich schema compensates.

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 description coverage is 91%, so the schema already documents nearly every parameter thoroughly. The description gestures at pipeline stages that map to rerank (CrossEncoder) and diverse (MMR) and mentions fetching full content via memory_get, but adds little syntax or semantics not already in the schema. Baseline 3 applies.

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?

States a specific verb and resource ('Search ALL memory: decisions, solutions, facts, lessons from ALL past sessions') and explicitly names siblings it is not (memory_search_fast, memory_explain_search). An agent can distinguish this from other recall tools without opening a 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?

Gives a clear directive ('Use BEFORE starting any task') and routes to alternatives ('Use memory_search_fast / memory_explain_search for explicit fast routing'). It stops short of stating when NOT to use this tool or how it differs from memory_recall_iterative / memory_temporal_query, so no explicit exclusions.

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

Deploy Server

memory_recall

Instructions

Input Schema

Schema Changelog

TDQS

Other Tools