--- title: Ollama Generate API - Amazon Bedrock Models via the Ollama Interface description: Generate text for a single prompt with Amazon Bedrock models using the Ollama-compatible /api/generate endpoint. Streaming, images, structured output, and thinking. keywords: Ollama generate API, Ollama compatible API, Amazon Bedrock generate, Ollama /api/generate, streaming NDJSON, structured output, Ollama thinking --- # Generate API (Ollama Compatible) Generate text for a single prompt with Amazon Bedrock models through the Ollama `/api/generate` interface. Served under `/ollama` by default; the examples below use `$BASE`, which includes that prefix. ## At a glance - :material-swap-horizontal: **Drop-in Ollama compatibility** — Follows the Ollama `/api/generate` request and response shape, including its newline-delimited JSON streaming transport, so an existing Ollama client works by changing the base URL. - :material-text-box-outline: **Single-prompt simplicity** — A minimal shape for clients that send one prompt at a time rather than a full conversation. - :material-code-json: **Structured output** — `format` accepts `"json"` or a full JSON Schema, and the answer is constrained to it. - :material-cloud-lock: **Private AWS backend** — Served entirely by Amazon Bedrock models in your own AWS account — no traffic to third-party endpoints. - :material-image-multiple: **Image input beyond base64** — `images` also takes a URL, a data URI or an `s3://` URI on models that read images. - :material-swap-horizontal: **Differs from the Ollama API:** `raw`, `suffix`, `template` and `context` are refused with `400` and `context` is never returned; models are never resident, so `keep_alive` holds nothing loaded — see [Limits and behaviour to know](#limitations). !!! info "Base URL and route prefix" By default, all Ollama-compatible routes are prefixed with `/ollama`. This means the Generate API is available at `/ollama/api/generate` instead of `/api/generate`. You can customize this prefix using the `OLLAMA_ROUTES_PREFIX` configuration variable documented in [HTTP Server and MCP](operations_configuration_server.md#ollama-routes-prefix). The `curl` examples on this page use a `$BASE` variable that **must include this prefix** — set it to your scheme and host followed by `OLLAMA_ROUTES_PREFIX`: ```bash export BASE="https://your-host/ollama" # :// + OLLAMA_ROUTES_PREFIX ``` ```bash curl -X POST "$BASE/api/generate" \ -H "Authorization: Bearer $API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "amazon.nova-micro-v1:0", "prompt": "Why is the sky blue?", "stream": false }' ``` ## Endpoints { #available-endpoints } | Endpoint | Method | What It Does | Powered By | MCP Tool | |------------------|--------|-------------------------------------------------------|----------------------------|-------------------| | `/api/generate` | `POST` | Text generation for a single prompt, Ollama Generate API | Amazon Bedrock chat models | `ollama_generate` | `/api/generate` is served in full; new integrations usually pick [`/api/chat`](api_ollama_chat.md), which carries message history, tool calls and multi-turn thinking. ## Model Names Send the model names [`GET /api/tags`](api_ollama_models.md#get-apitags) publishes — those are the canonical identifiers this server resolves directly. A trailing `:latest` is accepted and stripped as a fallback when the exact name is not found. Short aliases accepted on this server's other APIs also work here even though `/api/tags` does not list them. A name learned from ollama.com (for example `llama3.2:3b`) is not available through this server and answers `404`. Every response echoes the model name exactly as the request spelled it, in `model`. **Find compatible models:** Call [`/search_models`](api_search_models.md) with `route=ollama_generate` to discover model IDs that support this route, or call [`GET /api/tags`](api_ollama_models.md#get-apitags) for the Ollama-shaped listing. ## Feature compatibility { #feature-compatibility }
| Feature | Status | Notes | |-----------------------------------|:----------------------------------------:|--------------------------------------------------------------------| | **Input** | | | | `prompt` | :material-check-circle:{ .success role="img" aria-label="Supported" } | Full support | | `system` | :material-check-circle:{ .success role="img" aria-label="Supported" } | Sent as the conversation's system instruction | | `images` | :material-cog:{ .model-dep role="img" aria-label="Model-dependent" } | Multimodal models only; base64, a URL, a data URI or an `s3://` URI | | `format` | :material-cog:{ .model-dep role="img" aria-label="Model-dependent" } | `"json"` or a JSON Schema object; same behavior as [`/api/chat`](api_ollama_chat.md#structured-output) | | `options` | :material-minus-circle:{ .partial role="img" aria-label="Partial" } | `temperature`, `top_p`, `top_k`, `seed`, `stop`, `num_predict` (max output tokens), `presence_penalty` and `frequency_penalty` are forwarded; runner options (`num_ctx`, `num_gpu`, `num_thread`, `num_batch`, `main_gpu`, `use_mmap`, `min_p`, and any other unknown key) are accepted and ignored | | `stream` | :material-check-circle:{ .success role="img" aria-label="Supported" } | Newline-delimited JSON; defaults to `true` — see [Streaming](api_ollama_chat.md#streaming) | | `think` | :material-check-circle:{ .success role="img" aria-label="Supported" } | Boolean, or `low`/`medium`/`high`/`max`; comes back in `thinking` | | `keep_alive` | :material-minus-circle:{ .partial role="img" aria-label="Partial" } | Ignored for residency — models are never resident — but still tells a request with no prompt apart as a load or an unload, [below](#loading-and-unloading) | | `raw`, `suffix`, `template`, `context` | :material-close-circle:{ .unsupported role="img" aria-label="Unsupported" } | Rejected with `400` — see [Fields Not Available](#fields-not-available) | | `logprobs` / `top_logprobs` | :material-close-circle:{ .unsupported role="img" aria-label="Unsupported" } | Rejected with `400` | | **Output** | | | | `response` | :material-check-circle:{ .success role="img" aria-label="Supported" } | Full support | | `thinking` | :material-cog:{ .model-dep role="img" aria-label="Model-dependent" } | Follows [`CHAT_COMPLETIONS_REASONING_FIELD`](operations_configuration_observability.md#chat-completions-reasoning-field); omitted when the operator sets that to `none` | | `done_reason` | :material-check-circle:{ .success role="img" aria-label="Supported" } | `stop` or `length` on a generated answer; `load` or `unload` on a prompt-less request — see [Loading and Unloading](#loading-and-unloading) | | `context` | :material-close-circle:{ .unsupported role="img" aria-label="Unsupported" } | Never returned — see [Fields Not Available](#fields-not-available) | | **Usage tracking** | | | | `prompt_eval_count`, `eval_count` | :material-check-circle:{ .success role="img" aria-label="Supported" } | Real token counts | | `total_duration` | :material-check-circle:{ .success role="img" aria-label="Supported" } | Real wall-clock time | | `load_duration` | :material-close-circle:{ .unsupported role="img" aria-label="Unsupported" } | Never reported — there is no model-loading phase to measure | | `prompt_eval_duration`, `eval_duration` | :material-minus-circle:{ .partial role="img" aria-label="Partial" } | Reported only when streaming, measured from the stream itself; omitted on a non-streaming response |
**Legend:** * :material-check-circle:{ .success role="img" aria-label="Supported" } **Supported** — Fully compatible with the Ollama API * :material-cog:{ .model-dep role="img" aria-label="Model-dependent" } **Available on Select Models** — Check your model's capabilities * :material-minus-circle:{ .partial role="img" aria-label="Partial" } **Partial** — Supported with limitations * :material-close-circle:{ .unsupported role="img" aria-label="Unsupported" } **Unsupported** — Not available in this implementation
## Streaming By default (`stream` unset, or `true`), the response body is **newline-delimited JSON** (`application/x-ndjson`): one bare JSON object per line, with no `data:` prefix and no `[DONE]` sentinel. The stream is terminated by an object carrying `"done": true` and the response metrics. See [Streaming](api_ollama_chat.md#streaming) on the Chat API for the transport details and how a mid-stream failure is reported. ```bash curl -N -X POST "$BASE/api/generate" \ -H "Authorization: Bearer $API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "amazon.nova-micro-v1:0", "prompt": "Write a haiku about the sea." }' ``` ```json {"model":"amazon.nova-micro-v1:0","created_at":"2026-08-27T12:00:00.000000+00:00","response":"Waves","done":false} {"model":"amazon.nova-micro-v1:0","created_at":"2026-08-27T12:00:00.100000+00:00","response":" crash on silent shores","done":false} {"model":"amazon.nova-micro-v1:0","created_at":"2026-08-27T12:00:00.200000+00:00","response":"","done":true,"done_reason":"stop","total_duration":812345678,"prompt_eval_count":9,"prompt_eval_duration":123456789,"eval_count":11,"eval_duration":688888889} ``` ## Fields Not Available `raw`, `suffix`, `template` and `context` are each rejected with `400` when set. Every one of them needs the model's own prompt template or tokenizer to honor — the exact text the model was trained to expect around a prompt, or the token IDs of a prior turn — and neither is exposed here. The error names the offending field and points at [`/api/chat`](api_ollama_chat.md) as the alternative. Because there is no token context to hand back, `response.context` is never returned by this endpoint. Send `prompt` together with `system` for a system-prompted single turn, or use [`/api/chat`](api_ollama_chat.md) for a multi-turn conversation. ## Images `images` accepts base64-encoded image data, as Ollama does. This server additionally accepts a URL, a data URI or an `s3://` URI in the same field, on models that support image input. ## Loading and Unloading A request with **no `prompt`** is upstream's way of making a model resident — it is what `ollama run ` sends before it opens its REPL — and the same request with `keep_alive` set to `0` is what `ollama stop ` sends to evict it. A hosted model is always resident, so both are answered without invoking anything: ```json { "model": "amazon.nova-micro-v1:0", "created_at": "2026-01-01T00:00:00+00:00", "response": "", "done": true, "done_reason": "load" } ``` `done_reason` is `unload` when `keep_alive` is `0`, `load` otherwise. The answer is a single JSON object whatever `stream` says, as upstream's is. ## Limits and behaviour to know { #limitations } - `raw`, `suffix`, `template` and `context` are rejected with `400` — see [Fields Not Available](#fields-not-available). - `logprobs` and `top_logprobs` are rejected with `400` — log probabilities are not available. - `keep_alive` does not keep anything loaded: models are never resident. It is read only to tell a prompt-less request's `done_reason` apart, `load` from `unload`. - Runner options inside `options` (`num_ctx`, `num_gpu`, `num_thread`, `num_batch`, `main_gpu`, `use_mmap`, `min_p`, and any other key a local runner would use) are accepted and ignored. - `load_duration` is never reported: there is no model-loading phase to measure, and a number there would be invented. - `prompt_eval_duration` and `eval_duration` are reported only when streaming. All duration and count fields are optional in the Ollama API, so a client computing tokens-per-second from a non-streamed response has no duration to divide by. - A model name learned from ollama.com — `llama3.2:3b`, for one — names nothing this server serves and answers `404`; send a name [`GET /api/tags`](api_ollama_models.md#get-apitags) publishes. ## Request headers | Header | Purpose | Notes | |-----------------|-----------------|-------------------------------------------------| | `Authorization` | Gateway API key | `Bearer `, required like every other route | A local Ollama server needs no key; this one does, on every route. ## Try it ```bash curl -X POST "$BASE/api/generate" \ -H "Authorization: Bearer $API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "amazon.nova-micro-v1:0", "prompt": "Why is the sky blue?", "stream": false }' ``` **Example response:** ```json { "model": "amazon.nova-micro-v1:0", "created_at": "2026-08-27T12:00:00.000000+00:00", "response": "The sky is blue because of Rayleigh scattering...", "done": true, "done_reason": "stop", "total_duration": 734567890, "prompt_eval_count": 7, "eval_count": 42 } ``` ## Next steps Next: [Chat API](api_ollama_chat.md) · [Embed API](api_ollama_embed.md) · [Models API](api_ollama_models.md) · [Search Models API](api_search_models.md)