BroadRouter

Free LLM API tools

Updated

23 free tools for developers using LLM and speech-to-text APIs: token counters with each model's real tokenizer, cost, caching and flat-plan calculators, a context window checker, a rate-limit calculator, a JSON schema generator and a request builder. They run in your browser, need no sign-up and send nothing to a server.

Count and convert tokens

Real tokenizers for open models, not one-size estimates.

Token counters by model family, each running that family's own tokenizer: DeepSeek · GLM · Qwen · Gemma · GPT · Llama 3 · Llama 4 · Kimi · Mistral Small · Ministral 3 · gpt-oss · Phi-4 · Gemma 3 · MiniMax · Nemotron 3 · Qwen3.

Estimate and cut LLM costs

Published prices, cache hits, reasoning and flat plans, side by side.

Speech to text

Turn hours, calls or voice notes into a transcription bill.

Build and ship

Requests, schemas and limits for any OpenAI-compatible API.

Questions

Are these tools free?

Yes. Every tool is free with no sign-up and no usage limit. They run in your browser.

Is my text or data uploaded?

No. Calculations and token counting happen on your device. The token counters download a model's public tokenizer file once, then count locally. Shareable links keep your numbers in the page address, not on a server.

Where do the prices come from?

BroadNet Inference prices come from its published rate card (pricing.json), last synced 2026-11-14. Other providers' prices, such as the models in the LLM API cost calculator, are read from each provider's own pricing page with a source and check date per row (how we collect them). Where a calculator asks for your current provider's price, you can type it in.

Do the tools work with providers other than BroadNet?

Yes. The token counters, converters, schema generator, request builder and rate-limit calculator work for any OpenAI-compatible API. The cost calculators include a row for your current provider.

Can I link to or embed a tool?

Yes. Link to any tool page; most tools keep their inputs in the URL, so you can share a specific estimate.

Read as MarkdownAsk ChatGPTAsk ClaudeAsk Perplexity