Your next prompt.
One tool call away.

Search first, retrieve only what you need. Full original text, stable keys, source links, and preview videos come together.

In a browser agent

Open the library. Supported agents discover three read-only tools automatically:

search_loggedlol_prompts({"query":"animated hero", "videoOnly":true, "limit":5})
get_loggedlol_prompt({"key":"tower-floors"})
list_loggedlol_categories({})

Search covers complete prompt text as well as names and categories. Use category for an exact category, quoted phrases for exact text, and -term to exclude a term. Follow nextOffset for more results (maximum 50 per call). Tools use the current collection, including an imported collection.

Fetch images from jsDelivr

Use the image catalog to find image URLs by prompt name. Each image URL uses https://cdn.jsdelivr.net/gh/NoseyGames/loggedlol-media@COMMIT/assets/FILENAME. Images support browser fetch without authentication. The correct hostname is cdn.jsdelivr.net, not cdn.jsdelivr.com. Videos remain on GitHub raw URLs.

const catalog = await fetch("/api/v1/images.json").then(r => r.json());
const image = await fetch(catalog.items[0].url).then(r => r.blob());

Direct retrieval

Fetch the compact index, then follow a record’s json or text path. Each JSON record includes the complete prompt and media URLs. These are static resources; URL query parameters do not search.

GET /api/v1/index.json
GET /api/v1/prompts/tower-floors.json
GET /api/v1/prompts/tower-floors.txt

The Cloudflare Pages edition is public: fetch its JSON and text endpoints without authentication. Images on jsDelivr are public too. The older ChatGPT-hosted edition still requires its existing sign-in.

From the source checkout

Node.js 18+; no packages or credentials needed after cloning the private repo.

node agent.mjs search "animated hero"
node agent.mjs get tower-floors
node agent.mjs text tower-floors > prompt.txt

The local CLI searches the full collection and returns JSON to stdout, errors to stderr with a nonzero exit code. Static endpoints serve the published snapshot; browser tools reflect session imports. Rebuild static resources with node build-agent-api.mjs after scraping.

Data you can trust, instructions you choose