### Install all-the-public-replicate-models

Source: https://github.com/replicate/all-the-public-replicate-models/blob/main/README.md

Install the package using npm.

```sh
npm install all-the-public-replicate-models
```

--------------------------------

### Load Lite Model Metadata (JS)

Source: https://github.com/replicate/all-the-public-replicate-models/blob/main/README.md

Import and log basic public Replicate model metadata. This is a smaller subset, approximately 375KB.

```javascript
import models from 'all-the-public-replicate-models/lite'

console.log(models)
```

--------------------------------

### Load Full Model Metadata (JS)

Source: https://github.com/replicate/all-the-public-replicate-models/blob/main/README.md

Import and log all public Replicate model metadata. This includes approximately 17MB of data.

```javascript
import models from 'all-the-public-replicate-models'

console.log(models)
```

--------------------------------

### Load Model Stats (JS)

Source: https://github.com/replicate/all-the-public-replicate-models/blob/main/README.md

Import historical daily run counts for models and log the last 5 entries for a specific model. The stats are updated daily.

```javascript
import stats from 'all-the-public-replicate-models/stats'

console.log(stats["black-forest-labs/flux-schnell"].slice(-5))

/*
[ 
  { date: '2025-01-03', totalRuns: 176951005, dailyRuns: 1071498 },
  { date: '2025-01-04', totalRuns: 178025758, dailyRuns: 1074753 },
  { date: '2025-01-05', totalRuns: 179119496, dailyRuns: 1093738 },
  { date: '2025-01-06', totalRuns: 180272877, dailyRuns: 1153381 },
  { date: '2025-01-07', totalRuns: 181445133, dailyRuns: 1172256 }
]
*/
```

--------------------------------

### Dump All Model Metadata with CLI

Source: https://context7.com/replicate/all-the-public-replicate-models/llms.txt

Use the CLI aliases (`all-the-public-replicate-models`, `all-the-replicate-models`, `replicate-models`) to output all model metadata as JSON to stdout. This enables piping the data to other command-line tools for further processing.

```bash
# Dump all model metadata to stdout
npx all-the-public-replicate-models
```

```bash
# Save all model data to a local file
npx all-the-public-replicate-models > models.json
```

```bash
# Use replicate-models shorthand alias
npx replicate-models | jq 'length'
# 12000
```

--------------------------------

### Import Lite Model Metadata

Source: https://context7.com/replicate/all-the-public-replicate-models/llms.txt

Imports a lightweight array containing only the seven most useful fields per model. Find all text-to-image models and inspect the keys of a lite model.

```js
import models from 'all-the-public-replicate-models/lite';

// Find all text-to-image models
const textToImage = models.filter(m =>
  m.description?.toLowerCase().includes('text-to-image')
);

console.log(`Text-to-image models: ${textToImage.length}`);

// Each lite model has only 7 fields
const liteModel = models[0];
console.log(Object.keys(liteModel));
// [ 'url', 'owner', 'name', 'description', 'run_count', 'cover_image_url', 'github_url' ]
```

--------------------------------

### Import Historical Run Count Statistics

Source: https://context7.com/replicate/all-the-public-replicate-models/llms.txt

Imports a dictionary keyed by "owner/name" with daily run-count snapshots for the last 31 days. Look up a specific model's recent stats and find fastest-growing models.

```js
import stats from 'all-the-public-replicate-models/stats';

// Look up a specific model's recent stats
const fluxStats = stats["black-forest-labs/flux-schnell"];

// Get the last 5 days
console.log(fluxStats.slice(-5));
// [
//   { date: '2025-01-03', totalRuns: 176951005, dailyRuns: 1071498 },
//   { date: '2025-01-04', totalRuns: 178025758, dailyRuns: 1074753 },
//   { date: '2025-01-05', totalRuns: 179119496, dailyRuns: 1093738 },
//   { date: '2025-01-06', totalRuns: 180272877, dailyRuns: 1153381 },
//   { date: '2025-01-07', totalRuns: 181445133, dailyRuns: 1172256 }
// ]

// Find fastest-growing models over the past week
const recentGrowth = Object.entries(stats)
  .map(([nwo, history]) => {
    const lastWeek = history.slice(-7);
    const totalDailyRuns = lastWeek.reduce((sum, d) => sum + d.dailyRuns, 0);
    return { nwo, totalDailyRuns };
  })
  .sort((a, b) => b.totalDailyRuns - a.totalDailyRuns)
  .slice(0, 5);

console.log("Fastest growing models this week:", recentGrowth);
// [
//   { nwo: 'black-forest-labs/flux-schnell', totalDailyRuns: 7838124 },
//   { nwo: 'black-forest-labs/flux-dev',     totalDailyRuns: 2341000 },
//   ...
// ]
```

--------------------------------

### Save Model Metadata to File (CLI)

Source: https://github.com/replicate/all-the-public-replicate-models/blob/main/README.md

Use the CLI tool to redirect the output of all model metadata to a JSON file named 'models.json'.

```command
npx all-the-public-replicate-models > models.json
```

--------------------------------

### Import Full Model Metadata

Source: https://context7.com/replicate/all-the-public-replicate-models/llms.txt

Imports the complete array of all public Replicate models with full metadata. Inspect a single model's structure.

```js
import models from 'all-the-public-replicate-models';

// models is an array of model objects
console.log(`Total models: ${models.length}`);
// Total models: 12000+

// Inspect a single model's structure
const model = models[0];
console.log({
  owner: model.owner,               // e.g. "stability-ai"
  name: model.name,                 // e.g. "stable-diffusion"
  url: model.url,                   // e.g. "https://replicate.com/stability-ai/stable-diffusion"
  description: model.description,  // e.g. "A latent text-to-image diffusion model..."
  run_count: model.run_count,       // e.g. 52341829
  cover_image_url: model.cover_image_url,
  github_url: model.github_url,
  latest_version: model.latest_version, // Full OpenAPI schema, inputs/outputs
  default_example: model.default_example // A sample prediction object
});
```

--------------------------------

### Dump Model Metadata to JSON (CLI)

Source: https://github.com/replicate/all-the-public-replicate-models/blob/main/README.md

Use the CLI tool to output all model metadata as a JSON object to standard output.

```command
$ npx all-the-public-replicate-models
```

--------------------------------

### Filter and Sort Model Data with jq (CLI)

Source: https://github.com/replicate/all-the-public-replicate-models/blob/main/README.md

Pipe the CLI output to jq to filter for 'whisper' models, sort them by run count, and display their URL and run count.

```command
npx all-the-public-replicate-models | jq -r 'map(select(.name | contains("whisper"))) | sort_by(.run_count) | reverse | .[] | "\(.url) \(.run_count)"'
```

--------------------------------

### Filter Replicate Models with CLI and jq

Source: https://context7.com/replicate/all-the-public-replicate-models/llms.txt

Combine the CLI output with `jq` to filter and sort model data. This allows for complex queries directly in the shell, such as finding models by name, owner, or run count.

```bash
# Find all whisper models sorted by run count (descending)
npx all-the-public-replicate-models | jq -r '
  map(select(.name | contains("whisper")))
  | sort_by(.run_count)
  | reverse
  | .[] | "\(.url) \(.run_count)"
'
# https://replicate.com/openai/whisper 3790120
# https://replicate.com/m1guelpf/whisper-subtitles 38020
# https://replicate.com/hnesk/whisper-wordtimestamps 28889
```

```bash
# Extract only name + description for all models
npx all-the-public-replicate-models | jq '[.[] | {name: .name, owner: .owner, description: .description}]'
```

```bash
# Find models with over 1 million runs
npx all-the-public-replicate-models | jq '[.[] | select(.run_count > 1000000)] | length'
```

--------------------------------

### Find Top 10 Models by Run Count (JS)

Source: https://github.com/replicate/all-the-public-replicate-models/blob/main/README.md

Import all model metadata and use lodash-es to find and log the top 10 models sorted by their run count.

```javascript
import models from 'all-the-public-replicate-models'
import {chain} from 'lodash-es'

const mostRun = chain(models).orderBy('run_count', 'desc').take(10).value()
console.log({mostRun})
```

--------------------------------

### Query Replicate Models with lodash-es

Source: https://context7.com/replicate/all-the-public-replicate-models/llms.txt

Use lodash-es to chain operations like sorting, filtering, and mapping on the Replicate models array. This is useful for exploring the catalog and finding specific models based on criteria like run count or owner.

```javascript
import models from 'all-the-public-replicate-models';
import { chain } from 'lodash-es';

// Top 10 most-run models overall
const top10 = chain(models)
  .orderBy('run_count', 'desc')
  .take(10)
  .map(m => ({ url: m.url, run_count: m.run_count }))
  .value();

console.log("Top 10 models by run count:");
console.log(top10);
// [
//   { url: 'https://replicate.com/stability-ai/stable-diffusion', run_count: 52341829 },
//   ...
// ]

// Find all models owned by a specific organization
const blackForestModels = chain(models)
  .filter({ owner: 'black-forest-labs' })
  .orderBy('run_count', 'desc')
  .value();

console.log(`black-forest-labs has ${blackForestModels.length} public models`);
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.