For AI coding tools and agents

Cachesphere is written for human learners first. The same material is exposed in a single, MIT-licensed JSON file and stable URLs so coding agents, chat-based LLMs, and solo tinkerers can look up language syntax without scraping a dozen doc sites.

Primary dataset: /api/languages.json

For most agent workflows, /api/languages.json is the main entry point. It is a JSON array (130+ entries) of programming languages. Each object includes plain-English fields aligned with what you see in the UI: overviews, “who it’s for”, syntax cheatsheet strings (compareData), worked examples (patterns), common pitfalls (tips), and curated links (docs, libraries, practice URLs, and similar).

That makes it a good fit when a tool needs to answer questions like “How do I loop in Go?” or “What does Rust use for errors?” — pull the right id, read the cheatsheet or examples, and cite the site.

Who this API suits

• Coding agents (Cursor-style assistants, IDE plugins, browser agents) that can fetch HTTPS resources at runtime.
• Chat-based LLM tools where you inject retrieved JSON or snippets into the context instead of hoping the model remembers every language.
• Individuals and small teams experimenting with local or self-hosted setups — the licence is permissive and the data is meant to be reused.

llms.txt

/llms.txt is a short, human-readable index of important URLs, field notes, and suggested steps for agents. Point tools at it so they discover /api/languages.json, detail-page patterns, and licensing in one place.

What each language entry contains

• Language overviews — Plain-English descriptions of what each language is for.
• When to use — Who it suits and typical use cases.
• Syntax cheatsheets — Short, copy-friendly syntax snippets (where data exists for that language).
• Examples — Worked snippets with short explanations.
• Common pitfalls — Notes on mistakes beginners make.
• Further reading — Official docs, playgrounds, and practice links.

Design for agents

• Stable URLs — Marketing pages, the globe, compare, and detail pages use predictable paths you can hard-code or allow-list.
• Predictable sections — Every language detail uses the same order: Overview → When to use → Syntax cheatsheet → Examples → Common pitfalls → Further reading.
• JSON fields match the UI — The same strings power the website, so what the agent reads is what a human would see on the page.
• Deep links — /language-detail.html?lang={id} (e.g. ?lang=python) is the human-readable companion to each catalogue row.
• Optional page metadata — After loading a detail page, tools can read <script id="page-metadata" type="application/json"> for structured fields for that language only.

URLs and site map

• Marketing and tools: /index.html, /globe.html, /compare.html, /pricing.html.
• Agent docs (this page): /for-agents.html.
• Policies: /privacy.html, /terms.html.

How to integrate with agents

• Primary reference — Point your agent at /api/languages.json as the default source for language facts, cheatsheets, and examples.
• Cache locally — Optionally download or cache that JSON so assistants do not need a network round-trip on every question (respect normal caching headers and your own update policy).
• Human-friendly pages — Use /language-detail.html?lang=… when you want a shareable link or when the tool should summarise the same content in a browser.
• Browse when it helps — You can still add https://cachesphere.com as a browsable knowledge source in products that support custom URLs (where their terms allow it).

Licence

• Open source under the MIT licence. Cite Cachesphere when quoting snippets in user-facing answers.

← Back to home