+++
disableToc = false
title = "Model Gallery"
weight = 81
url = '/models'
+++

The model gallery is a curated collection of models configurations for [LocalAI](https://github.com/go-skynet/LocalAI) that enables one-click install of models directly from the LocalAI Web interface.

A list of the models available can also be browsed at [the Public LocalAI Gallery](https://models.localai.io).

LocalAI to ease out installations of models provide a way to preload models on start and downloading and installing them in runtime. You can install models manually by copying them over the `models` directory, or use the API or the Web interface to configure, download and verify the model assets for you. 


{{% notice note %}}
The models in this gallery are not directly maintained by LocalAI. If you find a model that is not working, please open an issue on the [main LocalAI repository](https://github.com/mudler/LocalAI/issues).
 {{% /notice %}}

{{% notice note %}}
GPT and text generation models might have a license which is not permissive for commercial use or might be questionable or without any license at all. Please check the model license before using it. The official gallery contains only open licensed models.
 {{% /notice %}}

![output](https://github.com/mudler/LocalAI/assets/2420543/7b16676e-d5b1-4c97-89bd-9fa5065c21ad)

## Instella-MoE availability

The gallery excludes `instella-moe-16b-a3b-think` and
`instella-moe-16b-a3b-think-q8` because the packaged llama.cpp backend does not
support their `instella-moe` architecture. Loading these GGUF files fails with
`unknown model architecture: 'instella-moe'`.

The entries can return after LocalAI ships a compatible backend. See
[the compatibility issue](https://github.com/mudler/LocalAI/issues/11681) and
[upstream llama.cpp support](https://github.com/ggml-org/llama.cpp/pull/26467).

## Useful Links and resources

- [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard) - here you can find a list of the most performing models on the Open LLM benchmark. Keep in mind models compatible with LocalAI must be quantized in the `gguf` format.

## How it works

Open **Models** in the WebUI. It is the canonical page for a model's complete
lifecycle and has two views:

- **Explore** browses configured galleries, compares hardware fit and variants,
  and installs models. This is the default view.
- **Installed** lists local model configurations and their running, idle,
  disabled, pinned, and distributed state. Select a model to load or stop it,
  edit its configuration, open a supported use case, inspect backend logs, or
  remove it.

Both views use the same model selection and store the view, search, filter, and
selection in the URL. Installing from Explore does not move you away from the
catalog; the entry updates in place when the operation finishes.

## NeoHorse-1-4B

Install NeoHorse-1-4B with automatic selection between Q4_K_M, Q6_K, and Q8_0 GGUF builds:

```bash
local-ai models install neohorse-1-4b-q4
```

To select Q8_0 explicitly:

```bash
local-ai models install neohorse-1-4b-q4 --variant neohorse-1-4b-q8
```

[NeoHorse-1-4B](https://huggingface.co/TokenRhythm/NeoHorse-1-4B) is a text-only Qwen3.5 fine-tune for coding, reasoning, and agentic tasks.
These builds use llama.cpp and the embedded Jinja chat template.
The gallery defaults to 32,768 context tokens; the model supports up to 262,144 tokens with sufficient memory.
The [GGUF downloads](https://huggingface.co/mradermacher/NeoHorse-1-4B-GGUF) are pinned to a revision and verified with SHA256 checksums.

## Spark-X2.5-1.7B

Install Spark-X2.5-1.7B with automatic selection between its Q4_K_M and Q8_0
GGUF builds:

```bash
local-ai models install spark-x2.5-1.7b-q4
```

To select the Q8_0 build explicitly:

```bash
local-ai models install spark-x2.5-1.7b-q4 --variant spark-x2.5-1.7b-q8
```

These text-only builds use the llama.cpp backend and the embedded Jinja chat
template. The gallery defaults to a 32,768-token context to limit memory use.
The [source model](https://huggingface.co/XHToken/Spark-X2.5-1.7B) supports up to
1,048,576 tokens; larger contexts require more memory. Use a current LocalAI
llama.cpp backend with Spark-X2.5 support.

## MiniCPM5-2B

MiniCPM5-2B offers the official Q4_K_M, Q8_0, and F16 GGUF builds through
llama.cpp. To install the F16 build explicitly:

```bash
local-ai models install minicpm5-2b --variant minicpm5-2b-f16
```

The F16 weights require a 5.04 GB download, plus additional memory for inference.
This entry uses the embedded chat template and an 8,192-token context.
See the [official GGUF repository](https://huggingface.co/openbmb/MiniCPM5-2B-GGUF).

## VRAM and download size estimates

When browsing the gallery or importing a model by URI, LocalAI can show **estimated download size** and **estimated VRAM** for models.

- **Where they appear**: In the model gallery table (Size / VRAM column), in the model detail modal, and after starting an import from URI (in the success message).
- **How they are computed**: GGUF models use file size (HTTP HEAD or local stat) and optional GGUF metadata (HTTP Range) for KV cache and overhead; other formats use Hugging Face file sizes and optional config when available. If metadata is unavailable, a size-only heuristic is used.
- **Hardware fit indicator**: When your system reports GPU or RAM capacity, the gallery shows whether the estimated VRAM fits (green) or may not fit (red) using a 95% headroom rule.
- Estimates are best-effort and may be missing if the server does not support HEAD/Range or the request times out.

## Gemma 4 Ortenzya 31B

[Ortenzya](https://huggingface.co/llmfan46/gemma-4-Ortenzya-The-Creative-Wordsmith-31B-it-uncensored-heretic-GGUF)
is a Gemma 4 fine-tune for creative writing, translation, and roleplay.
Install it with automatic selection between Q4_K_M and Q8_0:

```bash
local-ai models install gemma-4-31b-ortenzya-q4
```

To select Q8_0 explicitly:

```bash
local-ai models install gemma-4-31b-ortenzya-q4 --variant gemma-4-31b-ortenzya-q8
```

Both builds use llama.cpp, the embedded Jinja chat template, and the BF16
vision projector for image prompts. The gallery uses a 32,768-token context
to limit memory use, with temperature 1, top_p 0.95, and top_k 64.

## Gemma 4 12B IT

Install `gemma-4-12b-it-q4` for chat, tool use, and image prompts with
llama.cpp. The gallery offers Unsloth Q4_K_M, Q5_K_M, Q6_K, and Q8_0
builds, each with an F16 vision projector and the embedded chat template.
LocalAI selects a quantization based on available memory. To choose one
explicitly, run:

```bash
local-ai models install gemma-4-12b-it-q4 --variant gemma-4-12b-it-q8
```

These entries use a 32,768-token context and sampling defaults of
temperature 1, top_k 64, and top_p 0.95. They are separate from the
existing QAT builds. See the [source model](https://huggingface.co/google/gemma-4-12B-it)
and [GGUF files](https://huggingface.co/unsloth/gemma-4-12b-it-GGUF).

## Ornith 1.5 9B Uncensored

Junafinity's Ornith 1.5 9B Uncensored offers Q4_K_M, Q6_K, and Q8_0 GGUF
builds for llama.cpp. Each includes the F16 vision projector for image input.
Install with automatic variant selection:

```bash
local-ai models install ornith-1.5-9b-uncensored-q4
```

Select Q8 explicitly:

```bash
local-ai models install ornith-1.5-9b-uncensored-q4 --variant ornith-1.5-9b-uncensored-q8
```

These entries use the embedded chat template and a 32,768-token context.
Sampling defaults are temperature 0.6, top_p 0.95, top_k 20, min_p 0,
and repeat_penalty 1. The model has no MTP heads.
See the [source model](https://huggingface.co/junafinity/Ornith-1.5-9B-uncensored)
and [GGUF files](https://huggingface.co/mradermacher/Ornith-1.5-9B-uncensored-GGUF).

## Add other galleries

You can add other galleries by:

1. **Using the Web UI**: Navigate to the [Runtime Settings]({{%relref "features/runtime-settings#gallery-settings" %}}) page and configure galleries through the interface.

2. **Using Environment Variables**: Set the `GALLERIES` environment variable. The `GALLERIES` environment variable is a list of JSON objects, where each object has a `name` and a `url` field. The `name` field is the name of the gallery, and the `url` field is the URL of the gallery's index file, for example:

```json
GALLERIES=[{"name":"<GALLERY_NAME>", "url":"<GALLERY_URL"}]
```

3. **Using Configuration Files**: Add galleries to `runtime_settings.json` in the `LOCALAI_CONFIG_DIR` directory.

The models in the gallery will be automatically indexed and available for installation.

To use a gallery that needs authentication, such as a private GitHub repository or an internal server, add a matching entry to the credentials file. See [Private Registries and Galleries]({{% relref "advanced/private-sources" %}}).

## Dirk quantization variants

Dirk Qwen3.8 27B offers Q4_K_XL, Q5_K_XL, Q6_K_XL, and Q8_K_XL GGUF builds
with llama.cpp, MTP speculative decoding, and the shared F16 vision projector.
Select Q6 explicitly with:

```bash
local-ai models install dirk-qwen3.8-27b-q4 --variant dirk-qwen3.8-27b-q6
```

## Gallery mirrors

A gallery entry can declare a `mirrors` list of alternative locations for the same index file. Mirrors exist for availability, not for load balancing: LocalAI always prefers the `url`, and only falls back to the mirrors, in the order you listed them, when the one before it cannot be fetched. If the primary works, the mirrors are never contacted.

Mirrors accept any URI the gallery loader understands — `https://`, `github:`, `huggingface://` (also `hf://` and `hf.co/`), and `file://` — and the same rules apply to them as to a primary URL, so a `file://` mirror must still live inside your models directory.

```json
GALLERIES=[{"name":"localai", "url":"https://example.org/gallery/index.yaml", "mirrors":["github:mudler/LocalAI/gallery/index.yaml@master"]}]
```

Each attempt is bounded by a 120 second timeout, and a source that fails — a connection error, a timeout, or an HTTP error status such as 404 or 502 — is skipped for the next 10 minutes so a dead host is not re-dialled on every gallery listing. A source that answers is usable again immediately, and a request you cancel yourself is not counted against it. If every source happens to be inside that 10 minute window, LocalAI tries them all anyway rather than refuse to serve the gallery.

{{% notice warning %}}
**Neither mirrors nor the offline cache cover a `.ref` URL.** If a gallery's `url` ends in `.ref`, that reference file is fetched and resolved to the real index location *before* mirrors or the cached copy are consulted, and a failure to fetch it fails the gallery outright. That includes the offline case: a `.ref` gallery fails when the network is gone even if it has been fetched successfully before. Mirrors are alternates for the index, not for the reference that points at it. If you want mirror coverage or offline listings, point `url` directly at the index file.
{{% /notice %}}

The key is optional: a gallery without `mirrors` behaves exactly as before.

## Offline gallery listings

Every successful gallery fetch is written to a cache directory alongside your models directory (`<MODELS_PATH>/../cache/gallery/`), one file per gallery URL. If nothing can serve the index — the primary and every mirror failed, there is no network at all, the host is airgapped — LocalAI serves that last successfully fetched copy instead of failing the listing, and logs a warning saying it did so. This applies to every gallery whose `url` points directly at an index file, with or without `mirrors` — but not to a `.ref` URL, which is resolved before the cache is consulted (see the warning above).

Only a response that actually parses as a gallery index is stored. A captive portal, a proxy or a CDN can answer an index request with HTTP 200 and an HTML error page; caching that would replace a working offline copy with something no listing can read. An empty index is rejected for the same reason, so the previous copy survives.

Entries served this way may be stale: the copy is only as fresh as the last time the gallery could be reached, so models added or changed upstream since then will not show up, and an entry may point at a file that has since moved. A listing served from disk is a degraded mode, not a substitute for a reachable gallery.

The copy is deliberately kept out of the models directory itself, where LocalAI reads a `.yaml` file as an installed model's configuration. Deleting the cache directory is safe — the next successful fetch recreates it — and a machine that has never reached a gallery has nothing cached, so its first listing still fails.

## API Reference

### Model repositories

You can install a model in runtime, while the API is running and it is started already, or before starting the API by preloading the models.

To install a model in runtime you will need to use the `/models/apply` LocalAI API endpoint.

By default LocalAI is configured with the `localai` repository.

To use additional repositories you need to start `local-ai` with the `GALLERIES` environment variable:

```
GALLERIES=[{"name":"<GALLERY_NAME>", "url":"<GALLERY_URL"}]
```

For example, to spell out the default `localai` repository, you can start `local-ai` with:

```
GALLERIES=[{"name":"localai", "url":"https://index.localai.io/models", "mirrors":["github:mudler/LocalAI/gallery/index.yaml@master"]}]
```

`https://index.localai.io/models` is a caching mirror of the same index file, and the `github:` entry is the fallback used whenever it cannot be reached. `github:mudler/LocalAI/gallery/index.yaml@master` is expanded automatically to `https://raw.githubusercontent.com/mudler/LocalAI/master/gallery/index.yaml`.

Note: the url are expanded automatically for `github` and `huggingface`, however `https://` and `http://` prefix works as well.

#### Using Local Gallery Files

You can also use local gallery index files by using the `file://` prefix. For security reasons, **local gallery files must be located within your models directory** (the directory specified by `MODELS_PATH` or the default `models/` directory).

**Example:**

```json
GALLERIES=[{"name":"my-local-gallery", "url":"file:///path/to/models/my-gallery-index.yaml"}]
```

**Important notes:**
- The `file://` prefix is required for local paths
- The file path must be absolute (starting with `/` on Unix systems)
- The resolved path must be within your models directory for security
- If you try to access files outside the models directory, LocalAI will block the request

**Valid example** (assuming `MODELS_PATH=/opt/localai/models`):
```json
GALLERIES=[{"name":"local", "url":"file:///opt/localai/models/galleries/my-gallery.yaml"}]
```

**Invalid example** (file outside models directory):
```json
GALLERIES=[{"name":"local", "url":"file:///home/user/my-gallery.yaml"}]
```
This will be rejected with a security error.

{{% notice note %}}

If you want to build your own gallery, there is no documentation yet. However you can find the source of the default gallery in the [LocalAI repository](https://github.com/mudler/LocalAI/tree/master/gallery).
 {{% /notice %}}


### List Models

To list all the available models, use the `/models/available` endpoint:

```bash
curl http://localhost:8080/models/available
```

To search for a model, you can use `jq`:

```bash
curl http://localhost:8080/models/available | jq '.[] | select(.name | contains("replit"))'

curl http://localhost:8080/models/available | jq '.[] | .name | select(contains("localmodels"))'

curl http://localhost:8080/models/available | jq '.[] | .urls | select(. != null) | add | select(contains("orca"))'
```

### How to install a model from the repositories

Models can be installed by passing the full URL of the YAML config file, or either an identifier of the model in the gallery. The gallery is a repository of models that can be installed by passing the model name.

To install a model from the gallery repository, you can pass the model name in the `id` field. For instance, to install the `bert-embeddings` model, you can use the following command:

```bash
LOCALAI=http://localhost:8080
curl $LOCALAI/models/apply -H "Content-Type: application/json" -d '{
     "id": "localai@bert-embeddings"
   }'  
```

where:
- `localai` is the repository. It is optional and can be omitted. If the repository is omitted LocalAI will search the model by name in all the repositories. In the case the same model name is present in both galleries the first match wins.
- `bert-embeddings` is the model name in the gallery
  (read its [config here](https://github.com/mudler/LocalAI/tree/master/gallery/blob/main/bert-embeddings.yaml)).

### EfficientThink GGUF builds

[Qwen3.8-27B EfficientThink](https://huggingface.co/nerkyor/Qwen3.8-27B-EfficientThink-Uncensored-K3-Opus5-Grok4.6-GPT5.6Sol-SFT-SimPO-DFlash2-GGUF)
is available as Q6_K and Q8_0 builds for llama.cpp. Both include the matching
Q8 vision projector and use a 32,768-token context. The DFlash variants also
install the publisher's Q8 draft and enable speculative decoding.

To select a build explicitly:

```bash
local-ai models install --variant qwen3.8-27b-efficientthink-q8-dflash qwen3.8-27b-efficientthink-q6
```

Use `qwen3.8-27b-efficientthink-q6` or `qwen3.8-27b-efficientthink-q8` as the
variant name for ordinary decoding without a draft model.

### Model variants

The `qwen3.5-9b-defiant-fable-mtp` entry offers Q4_K_M and Q8_0 builds for
llama.cpp, with and without multi-token prediction (MTP). Each build includes
the BF16 vision projector. To select the Q8_0 MTP build explicitly:

```bash
local-ai models install localai@qwen3.5-9b-defiant-fable-mtp --variant qwen3.5-9b-defiant-fable-q8-mtp
```

Use `--variant qwen3.5-9b-defiant-fable-q8` for Q8_0 with ordinary decoding.

Some gallery entries offer several builds of the same model: different
quantizations, or the same weights served by a different engine. Such an entry
carries a `variants` list, and installing it normally lets LocalAI choose:

- variants whose backend cannot run on this machine are dropped;
- variants that do not fit the available memory are dropped. That budget is
  VRAM on a discrete-GPU host, and system RAM otherwise — including on
  unified-memory machines such as Apple Silicon, where the GPU shares system
  RAM and reports no separate VRAM pool;
- the entry's own build is never dropped. It competes with whatever survived
  rather than waiting for everything else to fail, so an entry that is itself
  the largest build that fits keeps its own payload;
- among the remaining builds the engine this machine prefers wins first: a vLLM
  build on an NVIDIA or AMD host, an MLX build on Apple Silicon, llama.cpp
  otherwise. The native accelerated runtime is worth more than a bigger
  download, so preference is settled before size;
- the largest build on the preferred engine then wins, because a bigger
  footprint means a higher quality build of the same model. A machine with no
  preferred engine picks purely by size;
- a build whose size could not be measured ranks below the entry's own build,
  so an unreadable size never quietly displaces the payload the entry ships;
- if nothing else survives, the entry's own build is installed. The entry is
  always installable, on any machine.

Because the entry's own build competes like every other candidate, the order of
the list means nothing and a `variants` list may offer smaller builds, larger
ones, or both.

Sizes are measured from the model's weights rather than downloaded, and cached.

The gallery listing only flags which entries offer variants, with a
`has_variants` field. It deliberately does not describe them: measuring a
variant is a network round trip per referenced build, so describing every
entry inline would make one listing request cost as many round trips as the
whole page has variants.

```bash
curl http://localhost:8080/api/models | jq '.models[] | select(.has_variants) | .name'
```

The Tiel-Coder gallery entry offers Q4, Q5, Q6, and Q8 MTP builds for
llama.cpp, each with the BF16 vision projector. To select the Q6 MTP build:

```bash
local-ai models install tiel-coder-35b-a3b-q4 --variant tiel-coder-35b-a3b-q6-mtp
```

### Collapsing the listing to one row per model

By default the listing returns every entry, including the individual builds a
parent entry offers as variants, so one model can occupy several rows. Pass
`collapse_variants=true` for the deduplicated view: every entry that is
installable in its own right, with nothing shown twice.

```bash
curl 'http://localhost:8080/api/models?collapse_variants=true'
```

An entry is hidden only when another entry already offers it as a variant, so
it stays reachable by installing that entry. Entries that declare variants are
always kept, and so is any entry nobody references. The filter is applied
before pagination, so page counts stay correct.

**Searching respects the collapse.** `term` is matched against every entry the
gallery holds, builds another entry offers included, so nothing becomes
unfindable. The collapse then decides how a match is reported: a hit on a build
another entry offers comes back as that entry, since that is the row installable
in its own right. Looking a model up by name must not answer "not found" for an
entry the gallery holds, and must not answer with a row the requested view has
no place for either. A term that is empty or only whitespace does not count as a
search:

```bash
# Collapsed: the parent stands in for the build it offers.
curl 'http://localhost:8080/api/models?collapse_variants=true'
# Collapsed, searching a build the parent offers: the parent comes back.
curl 'http://localhost:8080/api/models?collapse_variants=true&term=nanbeige4.1-3b-q8'
# Uncollapsed, same term: the build itself comes back.
curl 'http://localhost:8080/api/models?term=nanbeige4.1-3b-q8'
```

A parent appears once however many of its builds match, and once when it matches
in its own right as well. The substitution runs before the count and the page
math, so both describe the rows actually handed out.

`term`, `tag` and `backend` are all applied before the substitution, so each is
judged against the build that really carries the name, tag or backend rather
than against a parent that merely offers it. The consequence is worth knowing:
filtering by a backend only one variant declares returns that variant's parent,
whose own `backend` field may say something else. The alternative would be to
claim the gallery holds no such build.

The web UI requests the collapsed view by default and has a toggle for the other
one. The parameter stays on the API, off by default, for clients that want
either view.

Ask for the description one entry at a time, as the web UI does when you open
a model's variant menu:

```bash
curl http://localhost:8080/api/models/variants/localai@nanbeige4.1-3b-q4
```

```json
{
  "auto_selected": "nanbeige4.1-3b-q8",
  "variants": [
    { "model": "nanbeige4.1-3b-q8", "backend": "llama-cpp", "memory_bytes": 4187593113, "fits": true, "is_base": false },
    { "model": "nanbeige4.1-3b-q4", "backend": "llama-cpp", "fits": true, "is_base": true }
  ]
}
```

`auto_selected` is what installing without a choice would pick right now. `fits`
is whether auto-selection would consider that variant on this machine, and
`is_base` marks the entry's own build. `memory_bytes` is omitted entirely, as on
the second entry above, when the size could not be measured; read a missing
`memory_bytes` as unknown rather than as a free build.

An entry that declares no variants carries no `has_variants` field and answers
this endpoint with an empty list, so a client never has to ask about it.

To install a specific one, pass its name as `variant`:

```bash
curl $LOCALAI/models/apply -H "Content-Type: application/json" -d '{
     "id": "localai@nanbeige4.1-3b-q4",
     "variant": "nanbeige4.1-3b-q8"
   }'
```

An explicit choice is honored even when the machine looks too small for it, so
you can deliberately install a build LocalAI would not have picked. A `variant`
the entry does not declare fails the install and names what was requested; it
never quietly falls back to auto-selection. The choice is recorded, so a later
reinstall or upgrade of the same model stays on the variant you picked.

The same option exists on the CLI:

```bash
local-ai models install nanbeige4.1-3b-q4 --variant nanbeige4.1-3b-q8
```

Ornith-1.5-35B-A3B offers Q4_K_M, Q5_K_M, Q6_K, and Q8_0 GGUF builds
with a BF16 vision projector. For example, select Q6_K with:

```bash
local-ai models install ornith-1.5-35b-a3b-q4 --variant ornith-1.5-35b-a3b-q6
```

The `install_model` MCP tool takes the same `variant` argument, so an assistant
managing installs conversationally can pick a build too.

Entries without a `variants` list are unaffected by any of this and install
exactly as they always have.

### Artifact-backed models

Gallery models with an `artifacts` declaration are fully materialized during
installation. Their operation progresses through these phases:

```text
resolving -> downloading -> verifying -> committing -> persisting
```

The admin operations strip, the [Activity]({{% relref "operations/activity" %}})
page and `GET /api/operations` expose `currentBytes` and `totalBytes` as raw
transport bytes. Cancelling an active download leaves its partial files in
place so a retry can resume. A verification failure never
exposes a completed snapshot, while a retry or another installation reuses an
already verified content-addressed snapshot.

Deleting a model configuration does not delete its content-addressed snapshot
bytes. This allows another configuration or a later reinstall to reuse the
cache; safe cache garbage collection is deferred.

### How to install a model not part of a gallery

If you don't want to set any gallery repository, you can still install models by loading a model configuration file.

In the body of the request you must specify the model configuration file URL (`url`), optionally a name to install the model (`name`), extra files to install (`files`), and configuration overrides (`overrides`). When calling the API endpoint, LocalAI will download the models files and write the configuration to the folder used to store models.

```bash
LOCALAI=http://localhost:8080
curl $LOCALAI/models/apply -H "Content-Type: application/json" -d '{
     "config_url": "<MODEL_CONFIG_FILE_URL>"
   }' 
curl $LOCALAI/models/apply -H "Content-Type: application/json" -d '{
     "id": "<GALLERY>@<MODEL_NAME>"
   }' 
curl $LOCALAI/models/apply -H "Content-Type: application/json" -d '{
     "url": "<MODEL_CONFIG_FILE_URL>"
   }' 
```

An example that installs hermes-2-pro-mistral can be:
   
```bash
LOCALAI=http://localhost:8080
curl $LOCALAI/models/apply -H "Content-Type: application/json" -d '{
     "config_url": "https://raw.githubusercontent.com/mudler/LocalAI/v2.25.0/embedded/models/hermes-2-pro-mistral.yaml"
   }' 
```

The API will return a job `uuid` that you can use to track the job progress:
```
{"uuid":"1059474d-f4f9-11ed-8d99-c4cbe106d571","status":"http://localhost:8080/models/jobs/1059474d-f4f9-11ed-8d99-c4cbe106d571"}
```

For instance, a small example bash script that waits a job to complete can be (requires `jq`):

```bash
response=$(curl -s http://localhost:8080/models/apply -H "Content-Type: application/json" -d '{"url": "$model_url"}')

job_id=$(echo "$response" | jq -r '.uuid')

while [ "$(curl -s http://localhost:8080/models/jobs/"$job_id" | jq -r '.processed')" != "true" ]; do 
  sleep 1
done

echo "Job completed"
```

To preload models on start instead you can use the `PRELOAD_MODELS` environment variable.

<details>

To preload models on start, use the `PRELOAD_MODELS` environment variable by setting it to a JSON array of model uri:

```bash
PRELOAD_MODELS='[{"url": "<MODEL_URL>"}]'
```

The value must be a JSON array. Boolean values such as `true` are not valid model preload configurations.

Note: `url` or `id` must be specified. `url` is used to a url to a model gallery configuration, while an `id` is used to refer to models inside repositories. If both are specified, the `id` will be used.

For example:

```bash
PRELOAD_MODELS=[{"url": "github:mudler/LocalAI/gallery/stablediffusion.yaml@master"}]
```

or as arg:

```bash
local-ai --preload-models '[{"url": "github:mudler/LocalAI/gallery/stablediffusion.yaml@master"}]'
```

or in a YAML file:

```bash
local-ai --preload-models-config "/path/to/yaml"
```

YAML:
```yaml
- url: github:mudler/LocalAI/gallery/stablediffusion.yaml@master
```

</details>

{{% notice note %}}

You can find already some open licensed models in the [LocalAI gallery](https://github.com/mudler/LocalAI/tree/master/gallery).

If you don't find the model in the gallery you can try to use the "base" model and provide an URL to LocalAI:

<details>

```
curl $LOCALAI/models/apply -H "Content-Type: application/json" -d '{
     "url": "github:mudler/LocalAI/gallery/base.yaml@master",
     "name": "model-name",
     "files": [
        {
            "uri": "<URL>",
            "sha256": "<SHA>",
            "filename": "model"
        }
     ]
   }'
```

</details>

 {{% /notice %}}

### Override a model name

To install a model with a different name, specify a `name` parameter in the request body.

```bash
LOCALAI=http://localhost:8080
curl $LOCALAI/models/apply -H "Content-Type: application/json" -d '{
     "url": "<MODEL_CONFIG_FILE>",
     "name": "<MODEL_NAME>"
   }'  
```

For example, to install a model as `gpt-3.5-turbo`:
   
```bash
LOCALAI=http://localhost:8080
curl $LOCALAI/models/apply -H "Content-Type: application/json" -d '{
      "url": "github:mudler/LocalAI/gallery/gpt4all-j.yaml",
      "name": "gpt-3.5-turbo"
   }'  
```
### Additional Files

<details>

To download additional files with the model, use the `files` parameter:

```bash
LOCALAI=http://localhost:8080
curl $LOCALAI/models/apply -H "Content-Type: application/json" -d '{
     "url": "<MODEL_CONFIG_FILE>",
     "name": "<MODEL_NAME>",
     "files": [
        {
            "uri": "<additional_file_url>",
            "sha256": "<additional_file_hash>",
            "filename": "<additional_file_name>"
        }
     ]
   }'  
```

</details>

### Overriding configuration files

<details>

To override portions of the configuration file, such as the backend or the model file, use the `overrides` parameter:

```bash
LOCALAI=http://localhost:8080
curl $LOCALAI/models/apply -H "Content-Type: application/json" -d '{
     "url": "<MODEL_CONFIG_FILE>",
     "name": "<MODEL_NAME>",
     "overrides": {
        "backend": "llama",
        "f16": true,
        ...
     }
   }'  
```

</details>



## Examples

### Huihui Qwen3.8 Flash Next

Install the abliterated Qwen3.8-Flash-Next build for text chat and image input:

```bash
local-ai models install huihui-qwen3.8-flash-next-abliterated-q4
```

The entry downloads the publisher's four UD-Q4_K_XL GGUF shards and BF16
vision projector, approximately 112 GB in total. It uses llama.cpp with the
embedded chat template and a 32,768-token default context. The publisher
currently provides one quantization, so this entry has no alternative variants.

This model is modified to reduce refusals. See the
[model card](https://huggingface.co/huihui-ai/Huihui-Qwen3.8-Flash-Next-abliterated-GGUF)
and [Qwen Community License 1.0](https://huggingface.co/huihui-ai/Huihui-Qwen3.8-Flash-Next-abliterated-GGUF/blob/main/LICENSE)
for model details and usage terms.

### Embeddings: Bert

<details>

```bash
curl $LOCALAI/models/apply -H "Content-Type: application/json" -d '{
     "id": "bert-embeddings",
     "name": "text-embedding-ada-002"
   }'  
```

To test it:

```bash
LOCALAI=http://localhost:8080
curl $LOCALAI/v1/embeddings -H "Content-Type: application/json" -d '{
    "input": "Test",
    "model": "text-embedding-ada-002"
  }'
```

</details>

### Image generation: Stable diffusion

URL: https://github.com/EdVince/Stable-Diffusion-NCNN

{{< tabs >}}
{{% tab name="Prepare the model in runtime" %}}

While the API is running, you can install the model by using the `/models/apply` endpoint and point it to the `stablediffusion` model in the [models-gallery](https://github.com/mudler/LocalAI/tree/master/gallery#image-generation-stable-diffusion):
```bash
curl $LOCALAI/models/apply -H "Content-Type: application/json" -d '{         
     "url": "github:mudler/LocalAI/gallery/stablediffusion.yaml@master"
   }'
```

{{% /tab %}}
{{% tab name="Automatically prepare the model before start" %}}

You can set the `PRELOAD_MODELS` environment variable:

```bash
PRELOAD_MODELS=[{"url": "github:mudler/LocalAI/gallery/stablediffusion.yaml@master"}]
```

or as arg:

```bash
local-ai --preload-models '[{"url": "github:mudler/LocalAI/gallery/stablediffusion.yaml@master"}]'
```

or in a YAML file:

```bash
local-ai --preload-models-config "/path/to/yaml"
```

YAML:
```yaml
- url: github:mudler/LocalAI/gallery/stablediffusion.yaml@master
```

{{% /tab %}}
{{< /tabs >}}

Test it:

```
curl $LOCALAI/v1/images/generations -H "Content-Type: application/json" -d '{
            "prompt": "floating hair, portrait, ((loli)), ((one girl)), cute face, hidden hands, asymmetrical bangs, beautiful detailed eyes, eye shadow, hair ornament, ribbons, bowties, buttons, pleated skirt, (((masterpiece))), ((best quality)), colorful|((part of the head)), ((((mutated hands and fingers)))), deformed, blurry, bad anatomy, disfigured, poorly drawn face, mutation, mutated, extra limb, ugly, poorly drawn hands, missing limb, blurry, floating limbs, disconnected limbs, malformed hands, blur, out of focus, long neck, long body, Octane renderer, lowres, bad anatomy, bad hands, text",
            "mode": 2,  "seed":9000,
            "size": "256x256", "n":2
}'
```

### Audio transcription: Whisper

URL: https://github.com/ggerganov/whisper.cpp

{{< tabs >}}
{{% tab name="Prepare the model in runtime" %}}

```bash
curl $LOCALAI/models/apply -H "Content-Type: application/json" -d '{         
     "url": "github:mudler/LocalAI/gallery/whisper-base.yaml@master",
     "name": "whisper-1"
   }'
```

{{% /tab %}}
{{% tab name="Automatically prepare the model before start" %}}

You can set the `PRELOAD_MODELS` environment variable:

```bash
PRELOAD_MODELS=[{"url": "github:mudler/LocalAI/gallery/whisper-base.yaml@master", "name": "whisper-1"}]
```

or as arg:

```bash
local-ai --preload-models '[{"url": "github:mudler/LocalAI/gallery/whisper-base.yaml@master", "name": "whisper-1"}]'
```

or in a YAML file:

```bash
local-ai --preload-models-config "/path/to/yaml"
```

YAML:
```yaml
- url: github:mudler/LocalAI/gallery/whisper-base.yaml@master
  name: whisper-1
```

{{% /tab %}}
{{< /tabs >}}

### Note

LocalAI will create a batch process that downloads the required files from a model definition and automatically reload itself to include the new model. 

Input: `url` or `id` (required), `name` (optional), `files` (optional)

```bash
curl http://localhost:8080/models/apply -H "Content-Type: application/json" -d '{
     "url": "<MODEL_DEFINITION_URL>",
     "id": "<GALLERY>@<MODEL_NAME>",
     "name": "<INSTALLED_MODEL_NAME>",
     "files": [
        {
            "uri": "<additional_file>",
            "sha256": "<additional_file_hash>",
            "filename": "<additional_file_name>"
        },
      "overrides": { "backend": "...", "f16": true }
     ]
   }
```

An optional, list of additional files can be specified to be downloaded within `files`. The `name` allows to override the model name. Finally it is possible to override the model config file with `override`.

The `url` is a full URL, or a github url (`github:org/repo/file.yaml`), or a local file (`file:///path/to/file.yaml`). 

{{% notice warning %}}
**Local file security restriction:** When using `file://` URLs, the file path must be within your models directory (specified by `MODELS_PATH`). Files outside this directory will be rejected for security reasons.
{{% /notice %}}

The `id` is a string in the form `<GALLERY>@<MODEL_NAME>`, where `<GALLERY>` is the name of the gallery, and `<MODEL_NAME>` is the name of the model in the gallery. Galleries can be specified during startup with the `GALLERIES` environment variable.

Returns an `uuid` and an `url` to follow up the state of the process:

```json
{ "uuid":"251475c9-f666-11ed-95e0-9a8a4480ac58", "status":"http://localhost:8080/models/jobs/251475c9-f666-11ed-95e0-9a8a4480ac58"}
```

To see a collection example of curated models definition files, see the [LocalAI repository](https://github.com/mudler/LocalAI/tree/master/gallery).

#### Get model job state `/models/jobs/<uid>`

This endpoint returns the state of the batch job associated to a model installation.

```bash
curl http://localhost:8080/models/jobs/<JOB_ID>
```

Returns a json containing the error, and if the job is being processed:

```json
{"error":null,"processed":true,"message":"completed"}
```

Installations are processed one at a time. A job submitted while another install
is still running is reported as queued until the installer picks it up:

```json
{"error":null,"processed":false,"message":"queued","phase":"queued"}
```

A job ID is queryable from the moment `/models/apply` returns it, so a `404`/`500`
from this endpoint means the ID is genuinely unknown rather than merely waiting
its turn.

### Genesis Hermes Final

Install the Qwen3.6-35B-A3B Genesis Hermes Final model for text chat,
function calling, and image input:

```bash
local-ai models install qwen3.6-35b-a3b-genesis-hermes-final
```

The gallery offers APEX, APEX Compact, and Q8_K_P builds for llama.cpp.
APEX and APEX Compact also have variants with multi-token prediction (MTP)
enabled. Each build includes the F16 vision projector and uses the embedded
Jinja chat template. LocalAI selects a variant according to available memory
and serving features. To select the plain APEX build explicitly:

```bash
local-ai models install --variant qwen3.6-35b-a3b-genesis-hermes-final qwen3.6-35b-a3b-genesis-hermes-final
```

These entries set a 131,072-token context, following the
[model card's guidance](https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-Final-GGUF)
for thinking mode. This context requires additional memory beyond the weights.
The uncensored model uses the Apache-2.0 license.