AI & LLM Tools
Helpers for working with large language models — count tokens and estimate API cost before you send a prompt, chunk documents for retrieval, compare embeddings with cosine similarity or project them to 2D. All computed locally in your browser.
-
LLM Token Counter
Estimate how many tokens your prompt uses and what an API call will cost. Live character, word and token counts with editable prices, all in-browser.
-
Cosine Similarity Calculator
Compute cosine similarity, dot product, Euclidean and Manhattan distance, and the angle between two vectors or embeddings. Runs in your browser.
-
LLM GPU Memory (VRAM) Calculator
Estimate the GPU memory (VRAM) an LLM needs for inference or training, from parameter count, precision (fp16/int8/int4) and optimizer. In-browser.
-
JSONL Converter & Validator
Convert between JSONL and a JSON array, and validate every line with exact line numbers for errors. Great for LLM fine-tuning datasets. In-browser.
-
LLM API Cost Calculator
Estimate and compare monthly LLM API costs from token usage and request volume, with editable per-model input/output prices. Runs in your browser.
-
AI Text Chunker
Split text into overlapping chunks for RAG and embeddings — by tokens, characters, words or sentences — with a live preview and JSON/JSONL export.
-
PII Redactor
Detect and mask emails, phone numbers, credit cards, SSNs, IPs and API keys in text, with reversible placeholders and a re-identify step. Runs locally.
-
LLM Tool-Call Schema Builder
Build valid function-calling JSON Schemas for the OpenAI and Anthropic APIs from a simple form, or infer one from a sample arguments object.
-
AI Chat Export Viewer
Open your ChatGPT or Claude data export (conversations.json), browse and search your chats, and export any conversation to Markdown or HTML. All in-browser.
-
Embedding Projector
Paste labelled embedding vectors and project them to 2D with PCA to see clusters, plus a cosine-similarity nearest-neighbour list. Runs in-browser.