Tool reference · Sessions & episodes
memory_episode_recall
Finds past episodes filtered by text, project, outcome, impact or concepts.
local stdio server read-only
When to use
- Before starting a task, you want to know how similar work went before.
- You want to list past failures on a topic so you do not repeat them.
- You want the most impactful breakthroughs for a project.
Parameters
| Name | Type | Default | Description |
|---|---|---|---|
query | string | — | Search narrative text |
project | string | — | — |
outcome | "breakthrough" | "failure" | "routine" | "discovery" | — | — |
min_impact | number | 0 | — |
concepts | string[] | — | — |
limit | integer | 10 | — |
Example
Arguments
{
"project": "my-api",
"outcome": "failure",
"concepts": [
"webhooks"
],
"limit": 5
}Result shape
{
"episodes": [
{
"id": "b71e2c90d4a84f6b8e1c2d3f4a5b6c7d",
"project": "my-api",
"timestamp": "2026-08-12T10:14:03Z",
"narrative": "Tried to retry webhooks from a cron job; it double-sent under load.",
"outcome": "failure",
"impact_score": 0.6,
"concepts": [
"webhooks",
"cron"
],
"key_insight": "Retries need a lock or a queue."
}
]
}query is a plain substring match on the narrative. With concepts, results are ranked by how many concepts overlap.
Values are illustrative; the keys follow the server's handler. MCP clients receive the result as JSON text content.
Server description
The description the server sends to your agent in tools/list, captured from the v14.7.0 source:
Find past episodes (experiences). Search by concepts, outcome, project, or impact.