---
title: "Embed what the multilingual index is missing"
method: POST
path: "/v1/workspaces/{workspaceId}/knowledge/vectors/backfill"
tags: ["Knowledge"]
---

# Embed what the multilingual index is missing

`POST /v1/workspaces/{workspaceId}/knowledge/vectors/backfill`

## Request

- Base URL: `https://shareout.site`
- URL: `https://shareout.site/v1/workspaces/{workspaceId}/knowledge/vectors/backfill`
- Auth: HTTP bearer

Owner/admin of the workspace, or an instance admin. Re-embeds the workspace's stored passages, notes and pages into `KNOWLEDGE_VECTORS_V2` with the configured model — only what is missing — as one paced Workflow (a second call within 30 minutes joins it). Without Workflows it runs a bounded pass inline; call again until `done`. Metered as `knowledge_embedding`. `reset` forgets what was written first (after recreating the index).

## Request body

- object
  - `reset` boolean

## Response `200`

Ran inline (no Workflows) — `{ started false, inline true, done, embedded }`

## Other responses

- `202`: Started — `{ started true, instance }`
- `403`: Access denied
- `501`: The multilingual index is not configured (`NOT_CONFIGURED`)

## Changes

- **2026-10-26** `22f0fad12c81`: 1 info
  - endpoint added

[Operation history](https://skmtc.dev/shareout/apis/shareout-rest-api/changes/v1/workspaces/:workspaceId/knowledge/vectors/backfill/post.md)

---

[API](https://skmtc.dev/shareout/apis/shareout-rest-api.md) - [All operations](https://skmtc.dev/shareout/apis/shareout-rest-api/llms.txt) - [OpenAPI document](https://skmtc.dev/shareout/apis/shareout-rest-api/revisions/1e3f7fd34f6d?raw)