cognee-mcp

Overview Schema Source Related Servers Discussions Admin

forget

Deletes specific data items, entire datasets, or all owned memory across relational, graph, and vector stores, with optional history cleanup.

Instructions

Delete data from memory.

Can target a single data item, a specific dataset (by name or id), or everything the user owns. Removes data from the relational DB, graph DB, and vector DB.

Input Schema

Name	Required	Description
`data_id`	No	UUID of a single data item to delete. Must be paired with `dataset` or `dataset_id` so the owning dataset is unambiguous. To find both, call `recall` with include_references=true and copy them from the hit to delete: from its `refs:` line, or the data_id from its `Evidence:` block and the dataset_id from its `dataset:` line.
`dataset`	No	Dataset name to delete entirely.
`dataset_id`	No	UUID of the dataset to delete entirely, or to scope `data_id`.
`everything`	No	If true, delete ALL data across all datasets.
`clean_history`	No	If true, also delete the user's own search and recall history (in every dataset) before the data is deleted; alone, with no target, delete only the history. Default false: forgetting one item keeps the history; forgetting a dataset or everything purges that data's history regardless.

Output Schema

Name	Required	Description	Default
`result`	Yes

Schema Changelog

Changes observed during successful MCP inspections.

Changed2 schema fields changedNov 2, 2026

addedInput schema / properties / clean_history

Added value: +{
+  "default": false,
+  "description": "If true, also delete the user's own search and recall history (in\nevery dataset) before the data is deleted; alone, with no target,\ndelete only the history. Default false: forgetting one item keeps the\nhistory; forgetting a dataset or everything purges that data's\nhistory regardless.",
+  "type": "boolean"
+}

changedInput schema / properties / data_id / description

Previous value: -"UUID of a single data item to delete. Must be paired with `dataset`\nor `dataset_id` so the owning dataset is unambiguous."New value: +"UUID of a single data item to delete. Must be paired with `dataset`\nor `dataset_id` so the owning dataset is unambiguous. To find both,\ncall `recall` with include_references=true and copy them from the\nhit to delete: from its `refs:` line, or the data_id from its\n`Evidence:` block and the dataset_id from its `dataset:` line."

Changed6 schema fields changedv1.6.0Sep 19, 2026
- addedInput schema / properties / data_id / anyOf
```
Added value: +[
+  {
+    "type": "string"
+  },
+  {
+    "type": "null"
+  }
+]
```
- removedInput schema / properties / data_id / type
```
Removed value: -"string"
```
- addedInput schema / properties / dataset / anyOf
```
Added value: +[
+  {
+    "type": "string"
+  },
+  {
+    "type": "null"
+  }
+]
```
- removedInput schema / properties / dataset / type
```
Removed value: -"string"
```
- addedInput schema / properties / dataset_id / anyOf
```
Added value: +[
+  {
+    "type": "string"
+  },
+  {
+    "type": "null"
+  }
+]
```
- removedInput schema / properties / dataset_id / type
```
Removed value: -"string"
```

Changed2 schema fields changedv1.5.1Aug 22, 2026

addedInput schema / properties / data_id

Added value: +{
+  "default": null,
+  "description": "UUID of a single data item to delete. Must be paired with `dataset`\nor `dataset_id` so the owning dataset is unambiguous.",
+  "type": "string"
+}

addedInput schema / properties / dataset_id

Added value: +{
+  "default": null,
+  "description": "UUID of the dataset to delete entirely, or to scope `data_id`.",
+  "type": "string"
+}

Changed7 schema fields changedv1.5.0Aug 19, 2026

addedInput schema / additionalProperties
```
Added value: +false
```
addedInput schema / properties / dataset / description
```
Added value: +"Dataset name to delete entirely."
```
removedInput schema / properties / dataset / title
```
Removed value: -"Dataset"
```

addedInput schema / properties / everything / description

Added value: +"If true, delete ALL data across all datasets."

removedInput schema / properties / everything / title
```
Removed value: -"Everything"
```
removedInput schema / title
```
Removed value: -"forgetArguments"
```

changedOutput schema / (root)

Previous value: -nullNew value: +{
+  "properties": {
+    "result": {
+      "items": {},
+      "type": "array"
+    }
+  },
+  "required": [
+    "result"
+  ],
+  "type": "object",
+  "x-fastmcp-wrap-result": true
+}

First observedv1.0.1Jun 17, 2026

TDQS

A4.2/5.0

Behavior4/5

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

With no annotations at all, the description carries the full load, and it does disclose blast radius: data is removed from the relational DB, graph DB, and vector DB simultaneously, so an agent knows this is irreversible, multi-store destruction rather than a soft flag. It does not state whether deletion can be undone, whether confirmation or elevated permissions are required, or what happens on partial failure.

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

Conciseness5/5

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

Two short sentences, front-loaded with the verb and resource, followed by the scope enumeration and the blast-radius disclosure. Every clause carries information; nothing is redundant with the title or repeats itself.

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 five-parameter destructive tool, the description covers purpose, target modes, and the multi-store effect, while parameters and return values are fully documented by the schema and output schema. The only material omission is the absence of any irreversibility or confirmation warning, which matters more here because no annotations signal destructiveness.

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 100%, including the data_id/dataset pairing and the clean_history history-purge rule, so the schema does all the semantic work. The description only restates the target categories the parameters already encode, adding no syntax, format, or edge-case detail beyond the schema. Baseline 3 is correct.

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 (delete) and resource (memory data), and explicitly enumerates the three targeting scopes: single item, dataset by name/id, or everything the user owns. The description also resolves the non-obvious tool name 'forget' into a concrete delete operation and names the storage backends affected, so an agent can distinguish it from the non-destructive siblings (recall, remember, code_search) immediately.

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?

Each invocation mode is spelled out ('Can target a single data item, a specific dataset (by name or id), or everything the user owns'), which tells an agent which mode maps to which situation. It stops short of any exclusion or caution — nothing says when not to delete or that deletion should be confirmed first. The pairing requirement for data_id lives in the schema rather than the description.

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

Deploy Server