# Appndo

> Appndo is a curated directory of apps for work and everyday life, plus skills and prompts for AI agents, and a daily-updated comparison of AI model API prices. Every listing is reviewed by an editor before it is published; descriptions, features and pricing are written from each product's official website.

- Site: https://www.appndo.com
- Text-first index for AI crawlers and agents: https://www.appndo.com/for-ai (published apps, skills, prompts and models, with descriptions and canonical links; bounded selections, not a complete export)
- Sitemap (every published page): https://www.appndo.com/sitemap.xml
- Contact: info@appndo.com

## Sections

- [Apps](https://www.appndo.com/apps): handpicked apps with what each one does, its main features and its pricing model (Free, Freemium or Paid). Detail pages live at /apps/{slug}. Category pages live at /apps/category/{category}, e.g. /apps/category/project-management; skills and prompts use /skills/category/{category} and /prompts/category/{category}.
- [Skills](https://www.appndo.com/skills): reusable skills for AI agents, synced from skills.sh, with install counts, an install command and, under an open license, the original SKILL.md. Detail pages live at /skills/{owner}/{repo}/{skill}; publishers at /skills/{owner}; skill sets at /skills/{owner}/{repo}.
- [Prompts](https://www.appndo.com/prompts): reusable prompts with fill-in variables and usage notes. Detail pages live at /prompts/{slug}.
- [Model pricing](https://www.appndo.com/models): API prices of language, image and video models from the AI labs, clouds (AWS Bedrock, Azure, Vertex AI and others) and hosts (OpenRouter, Together AI, fal and others), updated daily.
  - Language models: one page per model at /models/{maker}/{model}, e.g. /models/anthropic/claude-sonnet-4-5, with the official price, every seller's price per million input, output and cached tokens, context length and LMArena score.
  - Makers: /models/{maker}, e.g. /models/openai, lists every model a maker makes. Sellers: /models/providers/{provider}, e.g. /models/providers/openrouter, lists everything one seller offers.
  - Image models: https://www.appndo.com/models/image, priced per standard 1024×1024 image and at low, high and 4K quality where sellers list them. Video models: https://www.appndo.com/models/video, priced per second of video by resolution. Each image and video model also has a page at /models/{maker}/{model}, e.g. /models/google/gemini-3-pro-image.
  - Comparisons: /models/compare/{model-a}-vs-{model-b}, e.g. /models/compare/claude-opus-4-7-vs-gpt-5-5, for the best-scoring models on LMArena that have an official price; each states both prices, a sample monthly cost and both arena scores.
  - Prices are in USD and come from LiteLLM's model cost map and models.dev (both MIT), checked once a day; each page states the date. They are not live quotes.

## Notes

- Pricing labels are editorial summaries; check each app's official pricing page for current plans.
- Model API prices are collected daily from open datasets, not quoted by the sellers; cite the model page with its date and check the seller's pricing page for current terms.
- Skill install counts come from skills.sh and show the date they were last updated.
- App submissions: https://www.appndo.com/pricing

## Read-only public APIs

These endpoints return published content only. No authentication is required.

- GET https://www.appndo.com/api/public/search?q=design
  - Required `q`: search text (up to 100 characters).
  - Response: `items`, up to 6 matches per section, with `name`, `slug`, `description`, `kind`, and `href`. This is a bounded search, not a complete catalog export.
- GET https://www.appndo.com/api/public/models?category=official
  - Optional `category`: `official` (the model makers), `cloud`, `hosted` or `all` (default).
  - Response: `models`, every language model offer in that group, with `id` (the offer id; offers added from models.dev start with `md:{source}/`, followed by the seller's model id), `provider`, `category`, `input`, `output`, `cacheRead` and `cacheWrite` (USD per million tokens), `context`, `features`, `source` (the seller's pricing page) and `canonical` (the model's page at /models/{canonical}, when it has one).
- GET https://www.appndo.com/api/public/models?model=anthropic/claude-sonnet-4-5
  - Response: `offers`, every seller's price for one model; the value is the path after /models/ on its page.
  - Both answers are cached for 10 minutes and change at most once a day.
- GET https://www.appndo.com/api/public/models/v1/export
  - Every language model offer in one versioned answer (fields are only ever added to v1), including hidden offers and price changes held for review. Supports ETag / If-None-Match (304 when unchanged); cached for 10 minutes.
  - Top level: `version`, `generated_at`, `sources` (`litellm.commit` and `fetched_at`, `models_dev.fetched_at`), `licenses`, `units`, `offers`.
  - Each offer: `id`, `api_model` (the `model` value the seller's API takes), `api_model_aliases`, `provider`, `provider_name`, `category`, `canonical`, `canonical_name`, `data_source`, `source`, `context`, `max_input` (only when smaller than `context`), `max_output`, `features` (vision, reasoning, tools, web, promptCaching, structuredOutput), `deprecation_date` (a retirement date still ahead, or null), `prices`, `tiers`, `hidden`, `hidden_reason`, `held`.
  - `prices`: `input`, `output`, `cacheRead`, `cacheWrite5m`, `cacheWrite1h`, `batchInput`, `batchOutput` as decimal strings in USD per million tokens; `perRequest` (USD per request) and `webSearchPerQuery` (USD per query, medium search context; `webSearchPerQueryBySize` for each size). null means no price is given; "0" means free.
  - `tiers`: higher prices for long prompts, each with `above_tokens`, `inclusive` (true when the price starts at the threshold itself) and the five token prices.
  - `held`: a price change waiting for review, with `field`, `current` (still shown), `proposed`, `since` and `changes`.
  - Other paths under /api/public/models/v1/ answer 404 with JSON `{"error":"not_found"}`.
- GET https://www.appndo.com/api/public/apps?sort=name&offset=0
  - Optional `q`, `category` (exact category name), `sort` (`latest`, `name`, `updated`), and `offset` (nonnegative integer).
  - Response: `items` (up to 24 apps), `total`, and `nextOffset` (null when finished).
  - Use each item's `slug` for its canonical page at https://www.appndo.com/apps/{slug}.

Use canonical detail pages when citing Appndo. Pricing summaries are not live quotes; follow the official pricing source for current terms. Avoid repeated polling; follow pagination and back off on errors. Admin, account, and payment endpoints are not public Agent interfaces.