> ## Documentation Index > Fetch the complete documentation index at: https://docs.conare.ai/llms.txt > Use this file to discover all available pages before exploring further. > ## Agent Instructions > Conare's API is at https://api.conare.ai. Authenticate with `Authorization: Bearer $CONARE_API_KEY`. > The full OpenAPI spec is at https://api.conare.ai/openapi.yaml; each endpoint has a page under /api-reference, and appending .md to any page URL returns its Markdown. > Embed with model conare-embed-1; send input_type "query" for search queries. Rerank with conare-rerank-1. > Prefer the TypeScript SDK (`npm install conare`) or plain HTTP in examples. > Install the Conare skill with `npx skills add https://docs.conare.ai`; search these docs over MCP at https://docs.conare.ai/mcp; the whole site is at https://docs.conare.ai/llms-full.txt. # Embed texts > Returns one vector per input; set `type: query` for search queries, `document` for stored text. - A text is read up to its first 6,000 characters and 2,048 tokens; the rest is silently ignored, so split longer texts. - `output_dtype: int8` returns int8 values and one scale per vector (vector ≈ values × scale); cosine on the values is unchanged. ## OpenAPI ````yaml /openapi.yaml post /embed openapi: 3.1.0 info: title: Conare API version: '1' summary: Store your data, keep it in sync, and return the right context to your AI. description: > Write documents into namespaces, search them, and sync apps into them. - Base URL `https://api.conare.ai`; send `Authorization: Bearer sk-conare-…` (keys at conare.ai/dashboard/keys). - A multi-organization key sends `X-Conare-Org: ` on every request (`GET /organizations` lists them). - Bodies are JSON in snake_case; an unknown field is 400 `invalid_json` (`/v1` and `/v2` ignore it). Errors are `{"error": {"code", "message", "request_id"}}`. - Send every write with an `Idempotency-Key` and retry with the same key: a repeat within 24 hours replays the first answer. - 429 and 503 carry `Retry-After` (seconds). - `Conare-Degraded` marks a success that skipped or slowed part of search, e.g. `rerank` or `embed-busy`. contact: name: Conare url: https://conare.ai servers: - url: https://api.conare.ai security: - apiKey: [] tags: - name: Namespaces description: >- A namespace is a named set of documents, e.g. `docs` or `user-42`; the first write creates it. - name: Documents description: >- A document is `{ id, text, metadata }`, chunked and embedded on write; writing the same id replaces it. - name: Query description: >- Hybrid search (vector and keyword) over your namespaces, filtered by metadata and reranked. - name: Models description: >- Embed and rerank text with Conare's models, through Conare's own routes or OpenAI- and Cohere-compatible ones. - name: Sources description: Apps, buckets and custom sources, each kept in sync into one namespace. - name: Custom sources description: >- Documents you push from your own system, keyed by your `external_id`, with snapshot runs to mirror a whole set. - name: Parse description: Parse a file to Markdown without storing it. - name: Files description: >- Upload a file and Conare stores it as one document, each hit citing its page, slide, sheet rows or section. - name: Users description: >- Delete everything Conare holds of one of your users, across every namespace and source. - name: Events description: What changed, in order (`seq`), to read from a cursor. - name: Health description: Service health. - name: Organizations description: The organizations a key acts for. paths: /embed: post: tags: - Models summary: Embed texts description: > Returns one vector per input; set `type: query` for search queries, `document` for stored text. - A text is read up to its first 6,000 characters and 2,048 tokens; the rest is silently ignored, so split longer texts. - `output_dtype: int8` returns int8 values and one scale per vector (vector ≈ values × scale); cosine on the values is unchanged. operationId: embed requestBody: required: true content: application/json: schema: type: object required: - input properties: input: type: array minItems: 1 maxItems: 128 description: The texts to embed, up to 32 KiB each. items: type: string type: type: string enum: - document - query default: document description: '`query` for a search query, `document` for stored text.' output_dtype: type: string enum: - float - int8 default: float description: '`int8`: int8 values in [-127, 127] and one scale per vector.' example: input: - how do refunds work? type: query responses: '200': description: The vectors, in input order. headers: Conare-Degraded: $ref: '#/components/headers/ConareDegraded' Conare-Embed-Busy-Ms: $ref: '#/components/headers/ConareEmbedBusyMs' content: application/json: schema: type: object required: - embeddings - model - dimensions properties: embeddings: type: array description: 'Floats; with `output_dtype: int8`, integers.' items: type: array items: type: number scales: type: array description: >- With `output_dtype: int8`, one per vector (0 for an all-zero vector). items: type: number output_dtype: type: string enum: - int8 description: 'Present only with `output_dtype: int8`.' model: type: string dimensions: type: integer '400': $ref: '#/components/responses/Error' '401': $ref: '#/components/responses/Error' '410': $ref: '#/components/responses/Error' '429': $ref: '#/components/responses/Error' '503': $ref: '#/components/responses/Error' components: headers: ConareDegraded: description: > Set when a request succeeded without part of search, or slower; comma-separated values: - `rerank`: results not reranked. `rerank-partial`: the rerank partly finished under load; usable, and a retry adds load. - `embed`: no vector retrieval. `embed-busy`: waited on a full embedding queue (`Conare-Embed-Busy-Ms`). `embed-fallback`: a backup served identical vectors. - `keyword`: no keyword retrieval. `keyword-terms`: keyword retrieval read only a long query's first words. - `cursor`: a paged search's first page came without `next_cursor`; search again for the rest. - `answer`: the answer model was unavailable; results only. schema: type: string example: rerank,embed-fallback ConareEmbedBusyMs: description: >- With `Conare-Degraded: embed-busy`, the milliseconds spent waiting; on writes, lower your concurrency until it goes away. schema: type: integer minimum: 1 example: 8400 responses: Error: description: An error. `X-Request-Id` is also in the response headers. headers: X-Request-Id: schema: type: string Retry-After: description: Seconds to wait before retrying (429, 503). schema: type: integer content: application/json: schema: $ref: '#/components/schemas/Error' example: error: code: invalid_request message: text is required request_id: 7d0f4c1e-2b8a-4e39-9a61-3c5e8b2f9d04 schemas: Error: type: object required: - error properties: error: type: object required: - code - message - request_id properties: code: type: string description: > Stable, machine-readable, by status: - 400: `invalid_request`, `fetch_failed`, `invalid_json`, `invalid_namespace`, `invalid_id`, `invalid_metadata`, `invalid_filter`, `invalid_cursor`, `invalid`, `md5_mismatch`, `idempotency_key_required`, `invalid_path` - 401: `unauthorized`, `invalid_upload` - 402 (billed calls only): `credits_exhausted`, `spend_cap_reached`, `monthly_limit_reached`, `payment_failed` - 403: `forbidden`, `namespace_forbidden`, `key_scope_models_only`, `namespace_not_allowed` - 404: `namespace_not_found`, `document_not_found`, `source_not_found`, `model_not_found`, `key_not_found`, `file_not_found`, `parse_not_found`, `not_found` - 405: `method_not_allowed` - 409: `version_conflict`, `conflict`, `idempotency_in_progress`, `run_active`, `run_not_receiving`, `source_deleting`, `user_purging`, `upload_incomplete`, `namespace_has_sources`, `too_many_exclusions`, `push_source`, `pull_source`, `files_source`, `no_original`, `file_replaced`, `gcs_bucket_claimed`, `limit_reached` - 410: `upload_expired`, `model_retired` - 413: `payload_too_large`, `too_large` - 415: `unsupported_file`, `type_mismatch`, `unsupported_encoding` - 422: `idempotency_key_reused`, `no_text`, `unreadable_file`, `password_protected`, `gcs_access_denied`, `gcs_not_verified`, `gcs_bucket_not_found` - 429: `rate_limited` - 500: `internal_error`; 502: `upstream_error`; 501: `models_only` - 503: `unavailable`, `sources_unavailable`, `timeout`, `embed_unavailable`, `sync_storage_unavailable`, `sync_unconfigured`, `files_unconfigured`, `files_unavailable`, `gcs_unconfigured`, `gcs_unavailable`, `gcs_auth_failed` message: type: string request_id: type: string securitySchemes: apiKey: type: http scheme: bearer bearerFormat: sk-conare-… description: >- A key from conare.ai/dashboard/keys with full access, some namespaces, or models only (`/embed`, `/rerank`, `/v1`, `/v2/rerank`, `/parse`). ```` ## Related topics - [Embed texts (OpenAI-compatible)](https://docs.conare.ai/zh/api-reference/models/embed-texts-openai-compatible.md) - [TypeScript SDK](https://docs.conare.ai/zh/sdk.md) - [Read a file](https://docs.conare.ai/zh/api-reference/files/read-a-file.md) - [Get a parse result](https://docs.conare.ai/zh/api-reference/parse/get-a-parse-result.md) - [Parse a file to Markdown](https://docs.conare.ai/zh/api-reference/parse/parse-a-file-to-markdown.md) This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.