AI & LLM Engineering Toolkit

Estimate token counts and API costs for GPT-4o, Claude, Gemini, and Llama models offline

Create reusable prompt templates with named variables and a live fill-in preview. No API key required.

Split long Markdown documents into token-budget chunks for RAG pipelines and LLM context windows, keeping headings, code blocks, and tables intact — all offline in your browser.

Strip markdown fences, thinking tags, and chatty preamble from LLM/agent output to get clean JSON or plain text - instant, fully client-side.

Build JSON tool and function-call schemas for OpenAI, Anthropic and MCP from a visual form, with parameter types, enums and required flags validated as you type.

Assemble production-grade LLM system prompts in your browser - role, tone, audience, numbered constraints, output format, and few-shot examples composed into copy-paste text. No API needed.

Pack files into an LLM context window with per-file toggles, token estimates, and XML-tagged bundle output.

Estimate the GPU VRAM a local LLM needs: weights per quantization level, KV cache from layers, KV heads and context length, plus runtime overhead, checked against common GPUs and unified memory.

Scan pasted prompts and documents for prompt injection patterns — instruction overrides, role hijacks, jailbreak modes, system-prompt exfiltration, fake chat delimiters and invisible Unicode — with offline heuristics and per-match highlighting. Runs entirely in your browser.