Read and delete memories

GET /v1/memories lists the library, newest first, and GET /v1/memories/{memory_id} returns one memory with its status, text and original file. Both reads are free. DELETE /v1/memories/{memory_id} removes a memory from search and answers at once.

List and read#

An imported conversation has type: "conversation". Its first read page includes messages, with the original role, content and created_at plus 1-based turn, line_start and line_end. Later pages continue the readable transcript through text and next_cursor. Notes and ordinary documents have messages: null. Deleting a conversation snapshot uses the same endpoint and removes the facts derived from that source.

Request

curl "https://api.fryri.com/v1/memories?end_user_id=user_42&limit=2" \
  -H "Authorization: Bearer $FRYRI_API_KEY"

Response

{
  "memories": [
    {
      "id": "file_tjSa4DvE15aMARb4",
      "type": "document",
      "title": "Example Domain.html",
      "status": "ready",
      "status_reason": null,
      "created_at": "2026-09-25T09:03:01.562079Z",
      "source_ref": "https://example.com",
      "sha256": "ff67a9d764d6a2367a187734e697f6a53217db9a21c101d410a113ca871a299d"
    },
    {
      "id": "file_khWTQ9FQRD0vNdAx",
      "type": "text",
      "title": "Atlas budget",
      "status": "ready",
      "status_reason": null,
      "created_at": "2026-09-25T09:02:58.286297Z",
      "source_ref": null,
      "sha256": "c6a6c1b533130aa2c02bb580edb9263885ac229f1ebd7016e1a341885a908ee4"
    }
  ],
  "next_cursor": "eyJwIjp7ImkiOjQ1OSwiayI6..."
}

The list covers every note and document in the library, however it arrived. limit sets the page size. Default value is 20. Range [1, 100]. Pass next_cursor back as cursor for the next page; it is null on the last page. source_ref_prefix lists only the memories whose source_ref starts with it, such as one folder or codebase. sha256 is the SHA-256 of the content as it was added, for comparing with your own copy.

Request

curl "https://api.fryri.com/v1/memories/file_NkolfLrElMmM2fCn?end_user_id=user_42" \
  -H "Authorization: Bearer $FRYRI_API_KEY"

Response

{
  "id": "file_NkolfLrElMmM2fCn",
  "type": "text",
  "title": "Project Atlas notes",
  "status": "ready",
  "status_reason": null,
  "created_at": "2026-09-25T09:02:54.915644Z",
  "source_ref": "crm:note:881",
  "sha256": "886d5886b41216cf0dc8e2234f4ffb24cb5666425200e5e2c5ee33d841b91be4",
  "summary": "A brief note about Project Atlas, which uses Python and FastAPI, with a launch review scheduled for 14 October.",
  "text": "Project Atlas uses Python and FastAPI. The launch review is on 14 October.",
  "lines": {
    "start": 1,
    "end": 1,
    "total": 1
  },
  "next_cursor": null,
  "download_url": "https://****.r2.cloudflarestorage.com/****/files/****/****.txt?response-content-type=text%2Fplain%3B%20charset%3Dutf-8&X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=****%2F20260925%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Date=20260925T090639Z&X-Amz-Expires=300&X-Amz-SignedHeaders=host&X-Amz-Signature=****"
}

status is processing until search can find the memory, then ready. A failed memory carries a status_reason of processing_failed, unsupported_format or reading_allowance_reached. A memory that is still processing can carry queued: a memory added in the Fryri app waits for the daily document allowance. Memories added through the API never wait. A memory of an end user must be read with its end_user_id. Otherwise, the request fails with 404 not_found.

Read long text in parts#

text holds the text of the memory: a note as it was added, or the extracted text of a document. The text of a long document arrives in parts. lines says which lines a part holds (start and end, 1-based and inclusive, out of total), and next_cursor is set while more text follows. A search result's location.line_start counts the same lines.

Complete example

import os

import requests

API = "https://api.fryri.com/v1"
HEADERS = {"Authorization": f"Bearer {os.environ['FRYRI_API_KEY']}"}


def read_all(memory_id, **params):
    text, previous = "", None
    while True:
        # Send request
        response = requests.get(f"{API}/memories/{memory_id}", headers=HEADERS, params=params)
        response.raise_for_status()
        # Read response
        page = response.json()
        lines = page["lines"]
        if previous is not None:
            # A part that starts on the line where the last part stopped continues that line.
            text += "" if lines and lines["start"] == previous["end"] else "\n"
        text += page["text"] or ""
        if not page["next_cursor"]:
            return text
        previous, params["cursor"] = lines, page["next_cursor"]


print(read_all("file_NkolfLrElMmM2fCn", end_user_id="user_42"))

A cursor from another memory, or from an earlier version of its text, fails with 422 invalid_cursor.

Download the original#

download_url links to the original file for 5 minutes. Each read returns a fresh link, and download_url is null when no original file is available.

Delete#

Request

curl -X DELETE "https://api.fryri.com/v1/memories/file_khWTQ9FQRD0vNdAx?end_user_id=user_42" \
  -H "Authorization: Bearer $FRYRI_API_KEY"

Response

{
  "id": "file_khWTQ9FQRD0vNdAx",
  "deleted": true
}

A deleted memory stops appearing in search and answers at once, together with anything learned from it. A memory of an end user must be deleted with its end_user_id. Otherwise, the request fails with 404 not_found. A second delete of the same memory also fails with 404.

DELETE /v1/end-users/{end_user_id} erases everything stored for one end user in one call. See Manage your users.

Best Practices#

Reading

• Store the memory id next to your own record, so a later read or delete needs no search.
• Poll GET /v1/memories/{memory_id} every few seconds after an add, and stop at ready or failed.
• Follow next_cursor until it is null to read a long document in full.

Deleting

• Delete with the same end_user_id the memory was added with.
• Treat a 404 on a repeated delete as done: the memory is already gone.
• Call DELETE /v1/end-users/{end_user_id} from your own account-deletion flow.

Tip: A short note arrives in one part, so its first read returns next_cursor as null.

Warning: download_url stops working after 5 minutes. Read the memory again for a fresh link.