> ## Documentation Index
> Fetch the complete documentation index at: https://docs.deeda.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Local (on-device)

> 185 usable capability rows, 135 documented parameters.

What you can do with **Local (on-device)** through the harness, by feature
family: the capability rows you can rely on (linked to vendor docs) and
the **parameters** you can set for each. Configure parameters through the
workflow surfaces described in the [Workflow Schema](/harness/workflow-schema)
— model/turn/budget fields on `agent`, provider knobs under
`harness_config.sdk_settings.local`, and tool/sandbox policy at top level.

## How to read this page

Decision rules per capability state — what YOU (the reader/agent) should do:

| State badge            | Meaning                                                                                                                                                                                                                                                                             | Your action                                                                                                                                                                         |
| ---------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| *(none)* = implemented | Wired end-to-end through deeda primitives and live-probed.                                                                                                                                                                                                                          | Use freely.                                                                                                                                                                         |
| *model-dependent*      | Works only on a documented subset of the provider's models.                                                                                                                                                                                                                         | Pin a supporting `model` in your workflow; verify against the vendor docs link.                                                                                                     |
| *partial-compatible*   | One half of the contract is proven, the other is not (yet). Patterns: request shape accepted but execution unproven at scale; wired in deeda but live probe pending; vendor does it but no deeda knob exists; works on one dispatch path (e.g. CLI) but not another (e.g. raw API). | Usable — but treat the **Parameters table as the entire contract**. Do not assume vendor-parity beyond what is listed. If your task depends on the unproven half, test small first. |

Parameter table columns: **Parameter** is the provider-side path the harness
maps for you (you set it via workflow fields or `harness_config.sdk_settings.<provider>`
knobs — see [Runtimes](/harness/runtimes) for which knobs exist). **Default**
applies when you say nothing. **Allowed** is exhaustive — values outside it
fail validation. **Risk** is deeda's policy weight: `high` parameters
generally require elevated `tool_policy`/`sandbox_profile` and may trigger
human approval on hot-reload.

## Core generation

**Capabilities** (4/4 rows usable):

* `local.llamacpp.openai_compat` (llm.complete) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md#api-endpoints)
* `local.llamacpp.server` (llm.complete) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.ollama.generate` (llm.complete) — [vendor docs](https://docs.ollama.com/api/generate)
* `local.ollama.openai_compat` (llm.complete) — [vendor docs](https://ollama.com/blog/openai-compatibility)

**Parameters** (7):

| Parameter         | Type   | Default                       | Allowed | Risk   | Notes                                                                            |
| ----------------- | ------ | ----------------------------- | ------- | ------ | -------------------------------------------------------------------------------- |
| `base_url`        | string | `"http://127.0.0.1:8080/v1"`  | —       | medium | [docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md) |
| `model`           | string | `"gemma-4-e2b-it"`            | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md) |
| `base_url`        | string | `"http://127.0.0.1:8080"`     | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md) |
| `base_url`        | string | `"http://127.0.0.1:11434"`    | —       | medium | [docs](https://docs.ollama.com/api)                                              |
| `model`           | string | —                             | —       | low    | Discovered via local.ollama.tags.                                                |
| `options.num_ctx` | number | `2048`                        | —       | low    | Context length window — model-dependent maximum.                                 |
| `base_url`        | string | `"http://127.0.0.1:11434/v1"` | —       | medium | [docs](https://ollama.com/blog/openai-compatibility)                             |

## llama-cpp-python

**Capabilities** (13/13 rows usable):

* `local.llamacpp.python.create_chat_completion` (llm.chat) — [vendor docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/)
* `local.llamacpp.python.create_completion` (llm.complete) — [vendor docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/)
* `local.llamacpp.python.create_embedding` (llm.embed) — [vendor docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/)
* `local.llamacpp.python.eval_sample_generate` (llm.streaming) — [vendor docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/)
* `local.llamacpp.python.from_pretrained` (provider.models) — [vendor docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/)
* `local.llamacpp.python.llama_cache_state` (provider.slots) — [vendor docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/)
* `local.llamacpp.python.llama_class` (provider.models) — [vendor docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/)
* `local.llamacpp.python.llama_grammar` (llm.structured\_output) — [vendor docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/)
* `local.llamacpp.python.logits_processor` (llm.sampling) — [vendor docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/)
* `local.llamacpp.python.save_load_state` (provider.slots) — [vendor docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/)
* `local.llamacpp.python.server_module` (provider.lifecycle) — [vendor docs](https://llama-cpp-python.readthedocs.io/en/latest/server/)
* `local.llamacpp.python.stopping_criteria` (llm.sampling) — [vendor docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/)
* `local.llamacpp.python.tokenize_detokenize` (llm.tokenize) — [vendor docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/)

**Parameters** (13):

| Parameter                             | Type   | Default | Allowed | Risk   | Notes                                                                    |
| ------------------------------------- | ------ | ------- | ------- | ------ | ------------------------------------------------------------------------ |
| `create_chat_completion.messages`     | array  | —       | —       | low    | [docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/) |
| `create_completion.prompt`            | string | —       | —       | low    | [docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/) |
| `create_embedding.input`              | string | —       | —       | low    | Requires an embedding GGUF (embedding=True).                             |
| `eval.tokens`                         | array  | —       | —       | low    | [docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/) |
| `from_pretrained.repo_id`             | string | —       | —       | medium | Live-probed with ggml-org/models tinyllamas/stories260K.gguf.            |
| `set_cache`                           | object | —       | —       | low    | [docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/) |
| `Llama.model_path`                    | string | —       | —       | low    | [docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/) |
| `create_completion.grammar`           | object | —       | —       | low    | GBNF; live-probed constraining output to yes\|no.                        |
| `create_completion.logits_processor`  | array  | —       | —       | medium | [docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/) |
| `save_state`                          | object | —       | —       | low    | [docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/) |
| `args.port`                           | number | —       | —       | medium | python -m llama\_cpp.server; OpenAI-compatible completions live-probed.  |
| `create_completion.stopping_criteria` | array  | —       | —       | low    | [docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/) |
| `tokenize.text`                       | string | —       | —       | low    | [docs](https://llama-cpp-python.readthedocs.io/en/latest/api-reference/) |

## llama.cpp CLI

**Capabilities** (4/4 rows usable):

* `local.llamacpp.bench` (eval.benchmark) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/README.md)
* `local.llamacpp.cli` (provider.cli) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/README.md)
* `local.llamacpp.model_acquisition_cache` (provider.models) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/README.md)
* `local.llamacpp.perplexity` (eval.perplexity) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/README.md)

**Parameters** (4):

| Parameter            | Type   | Default  | Allowed | Risk   | Notes                                                                           |
| -------------------- | ------ | -------- | ------- | ------ | ------------------------------------------------------------------------------- |
| `args.output_format` | string | `"json"` | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/blob/master/README.md)             |
| `args.prompt`        | string | —        | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/blob/master/README.md)             |
| `args.hf_repo`       | string | —        | —       | medium | -hf fetch + --cache-list; live-probed with ggml-org/SmolVLM-256M-Instruct-GGUF. |
| `args.file`          | string | —        | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/blob/master/README.md)             |

## llama.cpp CLI Tools

**Capabilities** (15/20 rows usable):

* `local.llamacpp.cli.batched` (eval.benchmark) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/batched)
* `local.llamacpp.cli.batched_bench` (eval.benchmark) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/batched-bench)
* `local.llamacpp.cli.embedding` (llm.embed) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/embedding)
* `local.llamacpp.cli.gguf_inspect` (provider.diagnostics) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/gguf)
* `local.llamacpp.cli.gguf_split` (file.split\_merge) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/gguf-split)
* `local.llamacpp.cli.imatrix` (provider.quantization) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/imatrix/README.md)
* `local.llamacpp.cli.lookahead` (llm.speculative\_decoding) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/lookahead)
* `local.llamacpp.cli.mtmd_cli` (media.multimodal) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/mtmd)
* `local.llamacpp.cli.parallel` (eval.benchmark) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/parallel)
* `local.llamacpp.cli.passkey` (eval.benchmark) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/passkey)
* `local.llamacpp.cli.quantize` (provider.quantization) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/quantize)
* `local.llamacpp.cli.retrieval` (llm.rag) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/retrieval)
* `local.llamacpp.cli.simple` (llm.complete) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/simple)
* `local.llamacpp.cli.simple_chat` (llm.chat) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/simple-chat)
* `local.llamacpp.cli.tokenize` (llm.tokenize) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/tokenize)

**Parameters** (15):

| Parameter                 | Type   | Default  | Allowed | Risk   | Notes                                                                             |
| ------------------------- | ------ | -------- | ------- | ------ | --------------------------------------------------------------------------------- |
| `args.npl`                | string | —        | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/batched-bench)     |
| `args.n_parallel`         | number | —        | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/batched)           |
| `args.embd_output_format` | string | `"json"` | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/embedding)         |
| `args.mode`               | string | `"r"`    | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/gguf)              |
| `args.split_max_size`     | string | `"50M"`  | —       | low    | Split + merge round trip live-probed.                                             |
| `args.train_file`         | string | —        | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/imatrix/README.md) |
| `args.n_predict`          | number | —        | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/lookahead)         |
| `args.image`              | string | —        | —       | low    | Vision CLI with -hf SmolVLM; live image description.                              |
| `args.n_sequences`        | number | —        | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/parallel)          |
| `args.junk`               | number | —        | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/passkey)           |
| `args.ftype`              | string | —        | —       | medium | [docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/quantize)          |
| `args.top_k`              | number | —        | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/retrieval)         |
| `args.ctx_size`           | number | —        | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/simple-chat)       |
| `args.n_predict`          | number | —        | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/simple)            |
| `args.prompt`             | string | —        | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/tokenize)          |

## llama.cpp Evaluation

**Capabilities** (1/1 rows usable):

* `local.llamacpp.eval.choice_logit_tasks` (eval.benchmark) · [*model-dependent*](#how-to-read-this-page) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/README.md)

## llama.cpp Quant Types

**Capabilities** (7/7 rows usable):

* `local.llamacpp.quant.bf16_f16_f32` (provider.quantization) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/quantize)
* `local.llamacpp.quant.iq_extreme` (provider.quantization) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/quantize)
* `local.llamacpp.quant.iq2` (provider.quantization) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/quantize)
* `local.llamacpp.quant.iq3` (provider.quantization) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/quantize)
* `local.llamacpp.quant.iq4` (provider.quantization) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/quantize)
* `local.llamacpp.quant.k_quants` (provider.quantization) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/quantize)
* `local.llamacpp.quant.legacy_q` (provider.quantization) — [vendor docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/quantize)

**Parameters** (7):

| Parameter    | Type   | Default     | Allowed | Risk   | Notes                                                                    |
| ------------ | ------ | ----------- | ------- | ------ | ------------------------------------------------------------------------ |
| `args.ftype` | string | `"f32"`     | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/quantize) |
| `args.ftype` | string | `"iq1_s"`   | —       | medium | Requires --imatrix; live-probed with a probe-built imatrix.              |
| `args.ftype` | string | `"iq2_xs"`  | —       | medium | Requires --imatrix; live-probed with a probe-built imatrix.              |
| `args.ftype` | string | `"iq3_xxs"` | —       | medium | Requires --imatrix; live-probed with a probe-built imatrix.              |
| `args.ftype` | string | `"iq4_xs"`  | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/quantize) |
| `args.ftype` | string | `"q4_K_M"`  | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/quantize) |
| `args.ftype` | string | `"q4_0"`    | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/tree/master/tools/quantize) |

## llama.cpp Runtime

**Capabilities** (4/7 rows usable):

* `local.llamacpp.runtime.cpu_memory` (provider.runtime) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/README.md)
* `local.llamacpp.runtime.kv_cache_context` (provider.context) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/README.md)
* `local.llamacpp.runtime.threads_batch` (provider.runtime) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/README.md)
* `local.llamacpp.sampling_controls` (llm.sampling) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/README.md)

**Parameters** (4):

| Parameter            | Type   | Default | Allowed | Risk | Notes                                                                      |
| -------------------- | ------ | ------- | ------- | ---- | -------------------------------------------------------------------------- |
| `args.threads`       | number | —       | —       | low  | -t/--no-mmap/--mlock live boot + completion.                               |
| `args.ctx_size`      | number | `2048`  | —       | low  | -c with -ctk/-ctv q8\_0; props echoes n\_ctx.                              |
| `args.threads_batch` | number | —       | —       | low  | system\_info echoes n\_threads\_batch.                                     |
| `sampling`           | object | —       | —       | low  | top\_k/top\_p/min\_p/temperature/penalties echoed in generation\_settings. |

## llama.cpp Server Anthropic Format

**Capabilities** (2/2 rows usable):

* `local.llamacpp.server.anthropic_count_tokens` (anthropic.messages.count\_tokens) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.anthropic_messages` (anthropic.messages) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)

**Parameters** (2):

| Parameter    | Type   | Default | Allowed | Risk | Notes                                                                            |
| ------------ | ------ | ------- | ------- | ---- | -------------------------------------------------------------------------------- |
| `messages`   | array  | —       | —       | low  | [docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md) |
| `max_tokens` | number | `1024`  | —       | low  | [docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md) |

## llama.cpp Server Native

**Capabilities** (21/26 rows usable):

* `local.llamacpp.server.apply_template` (llm.chat\_template) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.auth_tls` (provider.auth) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.completion_native` (llm.completion) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.detokenize` (llm.tokenize) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.embeddings_native` (llm.embedding) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.gpu_backend` (provider.gpu\_offload) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.grammar` (llm.structured\_output) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.health` (provider.health) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.infill` (llm.fim) · [*model-dependent*](#how-to-read-this-page) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.lora_adapters` (tuning.lora\_runtime) · [*model-dependent*](#how-to-read-this-page) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.metrics` (provider.metrics) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.parallel_batching` (provider.parallel\_decoding) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.props` (provider.runtime\_config) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.props_post` (provider.runtime\_config) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.reasoning` (llm.reasoning) · [*model-dependent*](#how-to-read-this-page) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.reranking` (llm.rerank) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.slots` (provider.slots) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.slots_save_restore` (provider.slots) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.speculative` (llm.speculative\_decoding) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.tokenize` (llm.tokenize) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.webui` (docs.webui) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)

**Parameters** (20):

| Parameter            | Type   | Default                   | Allowed | Risk   | Notes                                                                            |
| -------------------- | ------ | ------------------------- | ------- | ------ | -------------------------------------------------------------------------------- |
| `messages`           | array  | —                         | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md) |
| `args.api_key`       | string | —                         | —       | high   | 401 keyless / 200 bearer; HTTPS via --ssl-key-file/--ssl-cert-file live-probed.  |
| `base_url`           | string | `"http://127.0.0.1:8080"` | —       | medium | [docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md) |
| `n_predict`          | number | `128`                     | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md) |
| `tokens`             | array  | —                         | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md) |
| `content`            | string | —                         | —       | low    | Requires llama-server --embeddings.                                              |
| `args.n_gpu_layers`  | number | —                         | —       | low    | Metal backend on Apple Silicon; -ngl 99 live boot + completion.                  |
| `json_schema`        | object | —                         | —       | low    | GBNF --grammar twin; schema-constrained decoding live-probed.                    |
| `base_url`           | string | `"http://127.0.0.1:8080"` | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md) |
| `base_url`           | string | `"http://127.0.0.1:8080"` | —       | low    | Requires llama-server --metrics.                                                 |
| `messages.image_url` | object | —                         | —       | low    | Vision via mmproj; needs explicit --mmproj wiring on this build.                 |
| `args.n_parallel`    | number | `1`                       | —       | low    | props total\_slots reflects -np; concurrent completions live-probed.             |
| `body`               | object | —                         | —       | medium | Mutates global server properties; requires llama-server --props.                 |
| `base_url`           | string | `"http://127.0.0.1:8080"` | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md) |
| `query`              | string | —                         | —       | low    | Requires llama-server --reranking with an embedding/rerank model.                |
| `filename`           | string | —                         | —       | medium | Requires llama-server --slot-save-path; writes slot KV cache to disk.            |
| `base_url`           | string | `"http://127.0.0.1:8080"` | —       | low    | Requires llama-server --slots.                                                   |
| `args.model_draft`   | string | —                         | —       | medium | timings.draft\_n present in completions.                                         |
| `content`            | string | —                         | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md) |
| `base_url`           | string | `"http://127.0.0.1:8080"` | —       | low    | [docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md) |

## llama.cpp Server OpenAI Format

**Capabilities** (2/2 rows usable):

* `local.llamacpp.server.responses` (llm.responses) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)
* `local.llamacpp.server.tools` (tool.function\_calling) · [*model-dependent*](#how-to-read-this-page) — [vendor docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md)

**Parameters** (1):

| Parameter | Type   | Default | Allowed | Risk | Notes                                                                            |
| --------- | ------ | ------- | ------- | ---- | -------------------------------------------------------------------------------- |
| `input`   | string | —       | —       | low  | [docs](https://github.com/ggml-org/llama.cpp/blob/master/tools/server/README.md) |

## LocalAI Anthropic Format

**Capabilities** (1/1 rows usable):

* `local.localai.anthropic_messages` (anthropic.messages) — [vendor docs](https://localai.io/features/text-generation/)

## LocalAI Backends

**Capabilities** (1/23 rows usable):

* `local.localai.backend.llamacpp` (provider.backend) — [vendor docs](https://github.com/mudler/LocalAI/blob/master/README.md)

## LocalAI Galleries

**Capabilities** (2/6 rows usable):

* `local.localai.gallery_available` (provider.gallery) — [vendor docs](https://localai.io/docs/models/)
* `local.localai.gallery_jobs` (provider.gallery) — [vendor docs](https://localai.io/docs/models/)

## LocalAI Ollama Format

**Capabilities** (1/1 rows usable):

* `local.localai.ollama_compat` (llm.complete) — [vendor docs](https://github.com/mudler/LocalAI/blob/master/README.md)

## LocalAI OpenAI Format

**Capabilities** (3/16 rows usable):

* `local.localai.openai_chat` (llm.chat) — [vendor docs](https://localai.io/features/openai-functions/)
* `local.localai.openai_completions` (llm.completion) — [vendor docs](https://localai.io/features/text-generation/)
* `local.localai.openai_models` (provider.models) — [vendor docs](https://localai.io/features/text-generation/)

## MLX Apple Platform

**Capabilities** (2/2 rows usable):

* `local.mlx.platform.lazy_eval` (provider.runtime) — [vendor docs](https://ml-explore.github.io/mlx/build/html/usage/lazy_evaluation.html)
* `local.mlx.platform.macos_unified_memory` (provider.gpu\_offload) — [vendor docs](https://ml-explore.github.io/mlx/build/html/python/metal.html)

## MLX Distributed

**Capabilities** (2/2 rows usable):

* `local.mlx.distributed.launch` (provider.distributed\_inference) — [vendor docs](https://ml-explore.github.io/mlx/build/html/usage/distributed.html)
* `local.mlx.distributed.primitives` (provider.distributed\_inference) — [vendor docs](https://ml-explore.github.io/mlx/build/html/usage/distributed.html)

## MLX Examples

**Capabilities** (2/12 rows usable):

* `local.mlx.ex.bert` (llm.embed) · [*model-dependent*](#how-to-read-this-page) — [vendor docs](https://github.com/ml-explore/mlx-examples/tree/main/bert)
* `local.mlx.ex.t5` (llm.complete) · [*model-dependent*](#how-to-read-this-page) — [vendor docs](https://github.com/ml-explore/mlx-examples/tree/main/t5)

## MLX Fast Kernels

**Capabilities** (5/5 rows usable):

* `local.mlx.fast.metal_kernel` (provider.runtime) — [vendor docs](https://ml-explore.github.io/mlx/build/html/python/fast.html)
* `local.mlx.fast.quantized_matmul` (provider.quantization) — [vendor docs](https://ml-explore.github.io/mlx/build/html/python/fast.html)
* `local.mlx.fast.rms_norm` (provider.runtime) — [vendor docs](https://ml-explore.github.io/mlx/build/html/python/fast.html)
* `local.mlx.fast.rope` (provider.runtime) — [vendor docs](https://ml-explore.github.io/mlx/build/html/python/fast.html)
* `local.mlx.fast.scaled_dot_product_attention` (provider.attention\_backend) — [vendor docs](https://ml-explore.github.io/mlx/build/html/python/fast.html)

## MLX Optimizers

**Capabilities** (1/1 rows usable):

* `local.mlx.optimizers` (tuning.training\_reference) — [vendor docs](https://ml-explore.github.io/mlx/build/html/python/optimizers.html)

## MLX Profiling

**Capabilities** (3/3 rows usable):

* `local.mlx.metal.cache_limit` (provider.gpu\_offload) — [vendor docs](https://ml-explore.github.io/mlx/build/html/python/metal.html)
* `local.mlx.metal.capture` (provider.observability) — [vendor docs](https://ml-explore.github.io/mlx/build/html/python/metal.html)
* `local.mlx.utils.tree` (provider.runtime) — [vendor docs](https://ml-explore.github.io/mlx/build/html/python/tree_utils.html)

## MLX-LM CLI

**Capabilities** (10/17 rows usable):

* `local.mlx.lm.cli_benchmark` (eval.benchmark) — [vendor docs](https://github.com/ml-explore/mlx-lm/blob/main/mlx_lm/benchmark.py)
* `local.mlx.lm.cli_cache_prompt` (provider.slots) — [vendor docs](https://github.com/ml-explore/mlx-lm/blob/main/README.md)
* `local.mlx.lm.cli_chat` (llm.chat) — [vendor docs](https://github.com/ml-explore/mlx-lm/blob/main/README.md)
* `local.mlx.lm.cli_convert` (provider.models) — [vendor docs](https://github.com/ml-explore/mlx-lm/blob/main/README.md)
* `local.mlx.lm.cli_fuse` (tuning.lora\_runtime) — [vendor docs](https://github.com/ml-explore/mlx-lm/blob/main/mlx_lm/LORA.md)
* `local.mlx.lm.cli_generate` (llm.complete) — [vendor docs](https://github.com/ml-explore/mlx-lm/blob/main/README.md)
* `local.mlx.lm.cli_lora` (tuning.lora\_runtime) — [vendor docs](https://github.com/ml-explore/mlx-lm/blob/main/mlx_lm/LORA.md)
* `local.mlx.lm.cli_manage` (provider.models) — [vendor docs](https://github.com/ml-explore/mlx-lm/blob/main/mlx_lm/MANAGE.md)
* `local.mlx.lm.cli_perplexity` (eval.perplexity) · [*model-dependent*](#how-to-read-this-page) — [vendor docs](https://github.com/ml-explore/mlx-lm/blob/main/mlx_lm/perplexity.py)
* `local.mlx.lm.cli_server` (llm.chat) — [vendor docs](https://github.com/ml-explore/mlx-lm/blob/main/mlx_lm/SERVER.md)

## MLX-LM Python SDK

**Capabilities** (7/9 rows usable):

* `local.mlx.lm.kv_cache_quantized` (provider.kv\_cache) — [vendor docs](https://github.com/ml-explore/mlx-lm/blob/main/README.md)
* `local.mlx.lm.kv_cache_rotating` (provider.kv\_cache) — [vendor docs](https://github.com/ml-explore/mlx-lm/blob/main/README.md)
* `local.mlx.lm.py_generate` (llm.complete) — [vendor docs](https://github.com/ml-explore/mlx-lm/blob/main/README.md)
* `local.mlx.lm.py_load` (provider.models) — [vendor docs](https://github.com/ml-explore/mlx-lm/blob/main/README.md)
* `local.mlx.lm.py_prompt_cache` (provider.slots) — [vendor docs](https://github.com/ml-explore/mlx-lm/blob/main/README.md)
* `local.mlx.lm.py_sample_utils` (llm.sampling) — [vendor docs](https://github.com/ml-explore/mlx-lm/blob/main/mlx_lm/sample_utils.py)
* `local.mlx.lm.py_stream_generate` (llm.streaming) — [vendor docs](https://github.com/ml-explore/mlx-lm/blob/main/README.md)

## MLX-LM Server

**Capabilities** (1/1 rows usable):

* `local.mlx.lm.speculative` (llm.speculative\_decoding) — [vendor docs](https://github.com/ml-explore/mlx-lm/blob/main/mlx_lm/SERVER.md)

## Ollama Anthropic Format

**Capabilities** (1/3 rows usable):

* `local.ollama.anthropic_messages` (anthropic.messages) — [vendor docs](https://docs.ollama.com/api/anthropic-compatibility)

**Parameters** (1):

| Parameter    | Type   | Default | Allowed | Risk | Notes                                                       |
| ------------ | ------ | ------- | ------- | ---- | ----------------------------------------------------------- |
| `max_tokens` | number | `1024`  | —       | low  | [docs](https://docs.ollama.com/api/anthropic-compatibility) |

## Ollama Blobs

**Capabilities** (2/2 rows usable):

* `local.ollama.api_blobs_head` (file.exists) — [vendor docs](https://docs.ollama.com/api/blobs)
* `local.ollama.api_blobs_post` (file.upload) — [vendor docs](https://docs.ollama.com/api/blobs)

**Parameters** (2):

| Parameter | Type   | Default | Allowed | Risk   | Notes                                                               |
| --------- | ------ | ------- | ------- | ------ | ------------------------------------------------------------------- |
| `digest`  | string | —       | —       | low    | [docs](https://docs.ollama.com/api/blobs)                           |
| `digest`  | string | —       | —       | medium | Uploads content-addressed blobs (GGUF/adapter) to the local daemon. |

## Ollama CLI

**Capabilities** (6/7 rows usable):

* `local.ollama.cli_cp` (provider.models) — [vendor docs](https://docs.ollama.com/cli)
* `local.ollama.cli_pull_rm_ls` (provider.models) — [vendor docs](https://docs.ollama.com/cli)
* `local.ollama.cli_run` (llm.chat) — [vendor docs](https://docs.ollama.com/cli)
* `local.ollama.cli_serve` (provider.lifecycle) — [vendor docs](https://docs.ollama.com/cli)
* `local.ollama.cli_show` (provider.admin.read) — [vendor docs](https://docs.ollama.com/cli)
* `local.ollama.cli_stop` (provider.lifecycle) — [vendor docs](https://docs.ollama.com/cli)

**Parameters** (6):

| Parameter           | Type   | Default             | Allowed | Risk   | Notes                                                   |
| ------------------- | ------ | ------------------- | ------- | ------ | ------------------------------------------------------- |
| `args.source_dest`  | string | —                   | —       | low    | [docs](https://docs.ollama.com/cli)                     |
| `args.model`        | string | —                   | —       | medium | pull/rm/ls/create/cp management set.                    |
| `args.prompt`       | string | —                   | —       | low    | [docs](https://docs.ollama.com/cli)                     |
| `env.OLLAMA_HOST`   | string | `"127.0.0.1:11434"` | —       | medium | [docs](https://docs.ollama.com/cli)                     |
| `args.inspect_flag` | string | —                   | —       | low    | --modelfile/--parameters/--template/--system/--license. |
| `args.model`        | string | —                   | —       | low    | [docs](https://docs.ollama.com/cli)                     |

## Ollama Environment

**Capabilities** (13/15 rows usable):

* `local.ollama.env.context_length` (provider.context) — [vendor docs](https://docs.ollama.com/context-length)
* `local.ollama.env.debug` (provider.observability) — [vendor docs](https://docs.ollama.com/faq)
* `local.ollama.env.flash_attention` (provider.attention\_backend) — [vendor docs](https://docs.ollama.com/faq)
* `local.ollama.env.gpu_overhead` (provider.gpu\_offload) — [vendor docs](https://docs.ollama.com/faq)
* `local.ollama.env.host` (provider.connectivity) — [vendor docs](https://docs.ollama.com/faq)
* `local.ollama.env.keep_alive` (provider.lifecycle) — [vendor docs](https://docs.ollama.com/api/generate)
* `local.ollama.env.kv_cache_type` (provider.kv\_cache) — [vendor docs](https://docs.ollama.com/faq)
* `local.ollama.env.max_loaded` (provider.lifecycle) — [vendor docs](https://docs.ollama.com/faq)
* `local.ollama.env.max_queue` (provider.parallel\_decoding) — [vendor docs](https://docs.ollama.com/faq)
* `local.ollama.env.models_dir` (file.manage) — [vendor docs](https://docs.ollama.com/faq)
* `local.ollama.env.num_parallel` (provider.parallel\_decoding) — [vendor docs](https://docs.ollama.com/faq)
* `local.ollama.env.origins` (provider.connectivity) — [vendor docs](https://docs.ollama.com/faq)
* `local.ollama.env.sched_spread` (provider.gpu\_offload) — [vendor docs](https://docs.ollama.com/faq)

**Parameters** (13):

| Parameter                      | Type    | Default             | Allowed | Risk   | Notes                                                     |
| ------------------------------ | ------- | ------------------- | ------- | ------ | --------------------------------------------------------- |
| `env.OLLAMA_CONTEXT_LENGTH`    | number  | —                   | —       | low    | Live-probed: /api/ps reports the env-set context\_length. |
| `env.OLLAMA_DEBUG`             | boolean | —                   | —       | low    | Live-probed: DEBUG-level log lines under OLLAMA\_DEBUG=1. |
| `env.OLLAMA_FLASH_ATTENTION`   | boolean | —                   | —       | low    | Config echo + live generation under the flag.             |
| `env.OLLAMA_GPU_OVERHEAD`      | number  | —                   | —       | medium | Applied into scheduler config (startup config echo).      |
| `env.OLLAMA_HOST`              | string  | `"127.0.0.1:11434"` | —       | medium | [docs](https://docs.ollama.com/faq)                       |
| `env.OLLAMA_KEEP_ALIVE`        | string  | `"5m"`              | —       | low    | Live-probed: model evicted from /api/ps after expiry.     |
| `env.OLLAMA_KV_CACHE_TYPE`     | string  | —                   | —       | medium | Config echo + live generation under q8\_0.                |
| `env.OLLAMA_MAX_LOADED_MODELS` | number  | —                   | —       | low    | Applied into scheduler config (startup config echo).      |
| `env.OLLAMA_MAX_QUEUE`         | number  | —                   | —       | low    | Applied into scheduler config (startup config echo).      |
| `env.OLLAMA_MODELS`            | string  | —                   | —       | medium | [docs](https://docs.ollama.com/faq)                       |
| `env.OLLAMA_NUM_PARALLEL`      | number  | —                   | —       | low    | Applied into scheduler config (startup config echo).      |
| `env.OLLAMA_ORIGINS`           | string  | —                   | —       | medium | CORS allowlist; live-probed 403 deny / 200 allow.         |
| `env.OLLAMA_SCHED_SPREAD`      | boolean | —                   | —       | low    | Applied into scheduler config (startup config echo).      |

## Ollama Generate

**Capabilities** (6/6 rows usable):

* `local.ollama.chat` (llm.chat) — [vendor docs](https://docs.ollama.com/api/chat)
* `local.ollama.generate_context` (llm.state.continue) — [vendor docs](https://docs.ollama.com/api/generate)
* `local.ollama.generate_keep_alive` (provider.lifecycle) — [vendor docs](https://docs.ollama.com/api/generate)
* `local.ollama.generate_options_full` (llm.sampling) — [vendor docs](https://docs.ollama.com/modelfile)
* `local.ollama.generate_raw` (llm.complete) — [vendor docs](https://docs.ollama.com/api/generate)
* `local.ollama.generate_suffix` (llm.fim) · [*model-dependent*](#how-to-read-this-page) — [vendor docs](https://docs.ollama.com/api/generate)

**Parameters** (5):

| Parameter    | Type    | Default | Allowed | Risk   | Notes                                                                                                  |
| ------------ | ------- | ------- | ------- | ------ | ------------------------------------------------------------------------------------------------------ |
| `messages`   | array   | —       | —       | low    | [docs](https://docs.ollama.com/api/chat)                                                               |
| `context`    | array   | —       | —       | low    | [docs](https://docs.ollama.com/api/generate)                                                           |
| `keep_alive` | string  | `"5m"`  | —       | low    | [docs](https://docs.ollama.com/api/generate)                                                           |
| `options`    | object  | —       | —       | low    | Full sampler dict: temperature/top\_p/top\_k/repeat\_penalty/seed/num\_ctx/num\_predict (live-probed). |
| `raw`        | boolean | `false` | —       | medium | Bypasses the model prompt template.                                                                    |

## Ollama Model Management

**Capabilities** (11/11 rows usable):

* `local.ollama.api_create_adapters` (tuning.lora\_runtime) · [*model-dependent*](#how-to-read-this-page) — [vendor docs](https://docs.ollama.com/api/create)
* `local.ollama.api_create_quantize` (provider.quantization) — [vendor docs](https://docs.ollama.com/api/create)
* `local.ollama.api_create_safetensors` (provider.models) · [*model-dependent*](#how-to-read-this-page) — [vendor docs](https://docs.ollama.com/api/create)
* `local.ollama.api_show` (provider.admin.read) — [vendor docs](https://docs.ollama.com/api/show)
* `local.ollama.copy` (provider.models) — [vendor docs](https://docs.ollama.com/api/copy)
* `local.ollama.create` (provider.models) — [vendor docs](https://docs.ollama.com/api/create)
* `local.ollama.delete` (provider.models) — [vendor docs](https://docs.ollama.com/api/delete)
* `local.ollama.ps` (provider.lifecycle) — [vendor docs](https://docs.ollama.com/api/ps)
* `local.ollama.pull` (provider.models) — [vendor docs](https://docs.ollama.com/api/pull)
* `local.ollama.push` (provider.models) · [*model-dependent*](#how-to-read-this-page) — [vendor docs](https://docs.ollama.com/api/push)
* `local.ollama.tags` (provider.models) — [vendor docs](https://docs.ollama.com/api/tags)

**Parameters** (8):

| Parameter     | Type   | Default                    | Allowed | Risk   | Notes                                                                            |
| ------------- | ------ | -------------------------- | ------- | ------ | -------------------------------------------------------------------------------- |
| `quantize`    | string | `"q4_K_M"`                 | —       | medium | Requires an F16/F32 source tag; live-probed from smollm:135m-instruct-v0.2-fp16. |
| `model`       | string | —                          | —       | low    | [docs](https://docs.ollama.com/api/show)                                         |
| `destination` | string | —                          | —       | medium | [docs](https://docs.ollama.com/api/copy)                                         |
| `from`        | string | —                          | —       | medium | Medium risk: writes a new model manifest to the local store.                     |
| `model`       | string | —                          | —       | high   | High risk: destructive — removes a model from the local store.                   |
| `base_url`    | string | `"http://127.0.0.1:11434"` | —       | low    | [docs](https://docs.ollama.com/api/ps)                                           |
| `model`       | string | —                          | —       | medium | Medium risk: downloads model layers to local disk.                               |
| `base_url`    | string | `"http://127.0.0.1:11434"` | —       | low    | [docs](https://docs.ollama.com/api/tags)                                         |

## Ollama Modelfile

**Capabilities** (2/2 rows usable):

* `local.ollama.context_length` (provider.context) — [vendor docs](https://docs.ollama.com/context-length)
* `local.ollama.modelfile` (provider.modelfile) — [vendor docs](https://docs.ollama.com/modelfile)

**Parameters** (2):

| Parameter         | Type   | Default | Allowed | Risk   | Notes                                                                                   |
| ----------------- | ------ | ------- | ------- | ------ | --------------------------------------------------------------------------------------- |
| `options.num_ctx` | number | `4096`  | —       | low    | [docs](https://docs.ollama.com/context-length)                                          |
| `modelfile`       | object | —       | —       | medium | from/system/parameters/template create fields; live-probed via /api/create + /api/show. |

## Ollama OpenAI Format

**Capabilities** (6/6 rows usable):

* `local.ollama.openai_chat` (llm.chat) — [vendor docs](https://docs.ollama.com/api/openai-compatibility)
* `local.ollama.openai_completions` (llm.completions) — [vendor docs](https://docs.ollama.com/api/openai-compatibility)
* `local.ollama.openai_embeddings` (llm.embeddings) — [vendor docs](https://docs.ollama.com/api/openai-compatibility)
* `local.ollama.openai_images` (media.image\_generation) · [*model-dependent*](#how-to-read-this-page) — [vendor docs](https://docs.ollama.com/api/openai-compatibility)
* `local.ollama.openai_models` (provider.models) — [vendor docs](https://docs.ollama.com/api/openai-compatibility)
* `local.ollama.openai_responses` (llm.responses) — [vendor docs](https://docs.ollama.com/api/openai-compatibility)

**Parameters** (5):

| Parameter  | Type   | Default                       | Allowed | Risk | Notes                                                    |
| ---------- | ------ | ----------------------------- | ------- | ---- | -------------------------------------------------------- |
| `messages` | array  | —                             | —       | low  | [docs](https://docs.ollama.com/api/openai-compatibility) |
| `prompt`   | string | —                             | —       | low  | [docs](https://docs.ollama.com/api/openai-compatibility) |
| `input`    | string | —                             | —       | low  | Requires an embedding model (e.g. nomic-embed-text).     |
| `base_url` | string | `"http://127.0.0.1:11434/v1"` | —       | low  | [docs](https://docs.ollama.com/api/openai-compatibility) |
| `input`    | string | —                             | —       | low  | [docs](https://docs.ollama.com/api/openai-compatibility) |

## Ollama Server

**Capabilities** (1/1 rows usable):

* `local.ollama.api_version` (provider.health) — [vendor docs](https://docs.ollama.com/api/version)

**Parameters** (1):

| Parameter  | Type   | Default                    | Allowed | Risk | Notes                                       |
| ---------- | ------ | -------------------------- | ------- | ---- | ------------------------------------------- |
| `base_url` | string | `"http://127.0.0.1:11434"` | —       | low  | [docs](https://docs.ollama.com/api/version) |

## Ollama Streaming

**Capabilities** (1/1 rows usable):

* `local.ollama.streaming` (llm.streaming) — [vendor docs](https://docs.ollama.com/api/streaming)

**Parameters** (1):

| Parameter | Type    | Default | Allowed | Risk | Notes                                         |
| --------- | ------- | ------- | ------- | ---- | --------------------------------------------- |
| `stream`  | boolean | `true`  | —       | low  | [docs](https://docs.ollama.com/api/streaming) |

## Ollama Structured Output

> Concept guide: [Structured Output](/concepts/structured-output) — what this is, when to use it, and how to express it in workflow markdown.

**Capabilities** (1/1 rows usable):

* `local.ollama.structured_outputs` (llm.structured\_output) — [vendor docs](https://docs.ollama.com/capabilities/structured-outputs)

**Parameters** (1):

| Parameter | Type   | Default | Allowed | Risk | Notes                                                                             |
| --------- | ------ | ------- | ------- | ---- | --------------------------------------------------------------------------------- |
| `format`  | object | —       | —       | low  | JSON-schema constrained decoding; live-probed with a required-name object schema. |

## Ollama Thinking

> Concept guide: [Thinking & Reasoning](/concepts/thinking) — what this is, when to use it, and how to express it in workflow markdown.

**Capabilities** (1/1 rows usable):

* `local.ollama.thinking` (llm.thinking) · [*model-dependent*](#how-to-read-this-page) — [vendor docs](https://docs.ollama.com/capabilities/thinking)

## Ollama Vision

**Capabilities** (2/2 rows usable):

* `local.ollama.generate_image_input` (media.image\_input) · [*model-dependent*](#how-to-read-this-page) — [vendor docs](https://docs.ollama.com/api/generate)
* `local.ollama.vision` (media.image\_input) · [*model-dependent*](#how-to-read-this-page) — [vendor docs](https://docs.ollama.com/capabilities/vision)

## ollama-js SDK

**Capabilities** (3/5 rows usable):

* `local.ollama.js_abort_method` (llm.cancel) — [vendor docs](https://github.com/ollama/ollama-js)
* `local.ollama.js_async_iterator` (llm.streaming) — [vendor docs](https://github.com/ollama/ollama-js)
* `local.ollama.js_client_class` (provider.admin.read) — [vendor docs](https://github.com/ollama/ollama-js)

**Parameters** (3):

| Parameter     | Type    | Default | Allowed | Risk | Notes                                                              |
| ------------- | ------- | ------- | ------- | ---- | ------------------------------------------------------------------ |
| `abort`       | object  | —       | —       | low  | AbortError interrupts in-flight streamed generation (live-probed). |
| `chat.stream` | boolean | `true`  | —       | low  | [docs](https://github.com/ollama/ollama-js)                        |
| `Ollama.host` | string  | —       | —       | low  | [docs](https://github.com/ollama/ollama-js)                        |

## ollama-python SDK

**Capabilities** (10/11 rows usable):

* `local.ollama.python_async_client` (provider.admin.read) — [vendor docs](https://github.com/ollama/ollama-python)
* `local.ollama.python_chat_method` (llm.chat) — [vendor docs](https://github.com/ollama/ollama-python)
* `local.ollama.python_client_class` (provider.admin.read) — [vendor docs](https://github.com/ollama/ollama-python)
* `local.ollama.python_copy_delete` (provider.models) — [vendor docs](https://github.com/ollama/ollama-python)
* `local.ollama.python_create_modelfile` (provider.models) — [vendor docs](https://github.com/ollama/ollama-python)
* `local.ollama.python_embed_method` (llm.embed) — [vendor docs](https://github.com/ollama/ollama-python)
* `local.ollama.python_generate_method` (llm.complete) — [vendor docs](https://github.com/ollama/ollama-python)
* `local.ollama.python_list_method` (provider.models) — [vendor docs](https://github.com/ollama/ollama-python)
* `local.ollama.python_ps_method` (provider.lifecycle) — [vendor docs](https://github.com/ollama/ollama-python)
* `local.ollama.python_show_method` (provider.admin.read) — [vendor docs](https://github.com/ollama/ollama-python)

**Parameters** (10):

| Parameter          | Type   | Default | Allowed | Risk   | Notes                                           |
| ------------------ | ------ | ------- | ------- | ------ | ----------------------------------------------- |
| `AsyncClient.host` | string | —       | —       | low    | [docs](https://github.com/ollama/ollama-python) |
| `chat.messages`    | array  | —       | —       | low    | [docs](https://github.com/ollama/ollama-python) |
| `Client.host`      | string | —       | —       | low    | [docs](https://github.com/ollama/ollama-python) |
| `copy.source_dest` | string | —       | —       | medium | [docs](https://github.com/ollama/ollama-python) |
| `create.from_`     | string | —       | —       | medium | [docs](https://github.com/ollama/ollama-python) |
| `embed.input`      | string | —       | —       | low    | [docs](https://github.com/ollama/ollama-python) |
| `generate.prompt`  | string | —       | —       | low    | [docs](https://github.com/ollama/ollama-python) |
| `list`             | object | —       | —       | low    | [docs](https://github.com/ollama/ollama-python) |
| `ps`               | object | —       | —       | low    | [docs](https://github.com/ollama/ollama-python) |
| `show.model`       | string | —       | —       | low    | [docs](https://github.com/ollama/ollama-python) |

## Retrieval/files/embeddings

**Capabilities** (4/4 rows usable):

* `local.ollama.embed_dimensions` (llm.embed) — [vendor docs](https://docs.ollama.com/api/embed)
* `local.ollama.embed_truncate` (llm.embed) — [vendor docs](https://docs.ollama.com/api/embed)
* `local.ollama.embeddings` (llm.embed) — [vendor docs](https://docs.ollama.com/api/embed)
* `local.ollama.embeddings_legacy` (llm.embed) — [vendor docs](https://docs.ollama.com/api/embeddings)

**Parameters** (4):

| Parameter    | Type    | Default | Allowed | Risk | Notes                                                                                              |
| ------------ | ------- | ------- | ------- | ---- | -------------------------------------------------------------------------------------------------- |
| `dimensions` | number  | —       | —       | low  | Matryoshka truncation; honored by nomic-embed-text (live-probed at 64).                            |
| `truncate`   | boolean | `true`  | —       | low  | [docs](https://docs.ollama.com/api/embed)                                                          |
| `prompt`     | string  | —       | —       | low  | [docs](https://docs.ollama.com/api/embeddings)                                                     |
| `input`      | array   | —       | —       | low  | Requires an embedding model (e.g. nomic-embed-text); generation runners refuse embedding requests. |

## Tools

> Concept guide: [Tools & MCP](/concepts/tools-and-mcp) — what this is, when to use it, and how to express it in workflow markdown.

**Capabilities** (1/1 rows usable):

* `local.ollama.tools` (tool.call) · [*model-dependent*](#how-to-read-this-page) — [vendor docs](https://docs.ollama.com/capabilities/tool-calling)