Free LLM Navigator

What It Solves

Choosing free LLMs is tedious: quotas, context limits, multimodal support, stability, and free-tier policies change frequently. Developers usually need a quick answer to questions like “which models are currently available,” “which one fits Chinese chat, coding, or reasoning,” “how do I get the API key,” and “does it work after configuration.” This skill consolidates those steps into a conversational workflow, reducing back-and-forth across docs, signup pages, configs, and debugging.

How It Works

• Intent routing: When a user mentions “find free models,” “recommend a free API,” or similar intents, the skill triggers a search and recommendation flow.
• Information aggregation: It uses WebSearch to collect the latest free LLM API quota information and reads references/free_models_table.md to present model lists, scores, and scenario routing.
• Model operations: It supports commands such as compare, detail, apikey-guide, configure, test, recommend, and diagnose, covering comparison, signup guidance, configuration, connectivity checks, and troubleshooting.
• Scenario routing: It suggests primary and fallback models by use case, for example GLM-4-Flash for Chinese conversation, DeepSeek V3 for coding, DeepSeek R1 for reasoning or math, and Gemini 2.5 Pro for long documents or multimodal tasks.
• Switching execution: When a user explicitly asks to switch, it can call session_status(model="provider/model-id") to set the current session model and run an automatic test.

Boundaries

• Free quotas, signup rules, and model availability can change; results should be verified against live search and official pages.
• For expiring models, the skill estimates remaining days from the current conversation time, but it does not replace provider account status.
• This skill is intended for developers who need fast free-model selection and validation. It is not a production-only dependency and does not guarantee that a model remains free or consistently high-quality.