20 tools
AI & LLM Developer Tools
Twenty browser-based tools for building with large language models: count tokens, plan context and cost, lint and diff prompts, design structured outputs and tool schemas, work with MCP, and evaluate answers.
About these tools
These tools cover the engineering work around a model call rather than the call itself: how many tokens a prompt is, what a month of traffic costs, whether a system prompt contradicts itself, what schema a model must answer in, and how to tell whether one answer is better than another.
Almost everything runs in your browser. Prompts, responses, vectors and datasets you paste are processed in the tab and are never sent to a model, to analytics or to a2z.tools. Where a figure depends on a provider - prices, tokenizers - the page names the source and the date it was checked, and says plainly when a number is an estimate.
Popular tasks
All AI & LLM Developer Tools
- AI Token Counter Count tokens for OpenAI models with the real o200k and cl100k tokenizers, see labelled estimates for Claude and Gemini, and check how much of a context window your text uses.
- LLM API Cost Calculator Estimate LLM API spend from input, cached-input, output and batch tokens across OpenAI, Anthropic and Google, using a dated price catalogue you can override.
- Context Window Planner Allocate a model's context window between system prompt, tool definitions, retrieved chunks, history, user input and output reserve, and see exactly where it overflows.
- Prompt Template Builder Build reusable prompt templates with typed variables, defaults and conditional sections, render them against test cases, and export the template as JSON.
- System Prompt Linter Check a system prompt for contradictory instructions, priority conflicts, vague wording, repetition and a missing output contract, using deterministic rules and without sending it to any model.
- Prompt Version Diff Checker Compare two versions of a prompt section by section, see added and removed instructions and changed variables, and export a review report - all in your browser.
- Multi-Model Response Comparator Paste answers from several models and compare coverage of your key points, structure, claims, citations, readability, latency and cost you supply - no model is called.
- Structured Output Schema Builder Design the JSON Schema a model must answer in, field by field, validate example outputs against it, and export it in the shape each provider's structured-output feature expects.
- Function Calling Schema Builder Define a tool once - name, description, parameters, required fields and examples - validate it, and export the function or tool definition for OpenAI, Anthropic and Gemini.
- OpenAI / Anthropic / Gemini Request Converter Convert a chat request between OpenAI Chat Completions, Anthropic Messages and Gemini generateContent, with secrets redacted and every unsupported or changed field listed.
- MCP Server Configuration Builder Build the initialize-result capabilities, tool, resource and prompt declarations and client configuration for a Model Context Protocol server, validated against the current specification.
- MCP Capability Inspector Paste an MCP initialize, tools/list, resources/list or prompts/list response, or load public JSON from a URL, and see capabilities, schemas and compatibility issues - nothing is ever executed.
- RAG Chunking Simulator Split a document with fixed, sentence, paragraph or recursive chunking, tune size and overlap, and see token distribution and chunk boundaries before you embed anything.
- Embedding Similarity Explorer Compare embedding vectors with cosine similarity, dot product and Euclidean distance, check dimensions and normalisation, and rank candidates against a query vector.
- Vector Database Storage Calculator Estimate raw vector bytes, metadata, index overhead, replication and growth for a vector database, with every assumption visible and editable.
- AI Evaluation Dataset Builder Create evaluation cases with inputs, expected behaviour, tags and slices, validate them for duplicates and gaps, and export JSONL or CSV without anything leaving the page.
- LLM Evaluation Rubric Builder Define weighted criteria with anchored score levels and pass thresholds, preview deterministic scoring, and export the rubric and an LLM-as-judge prompt.
- Hallucination Claim Extractor Pull the checkable factual claims - numbers, dates, names, quotations, superlatives - out of AI-generated text so a person can verify them. It extracts; it does not decide truth.
- Citation Coverage Checker Map numbered or linked citations to the sentences around them, find factual statements with no citation, and catch missing, unused or out-of-order reference numbers.
- LLM Batch Request Builder Turn a list of prompts or CSV rows into valid batch files for the OpenAI Batch API or Anthropic Message Batches, with duplicate-ID checks and token and cost estimates.
Newly added
Related A2Z.Tools utilities
Other tool collections
Rate this tool
Was this tool useful? Your feedback helps us improve it.
No ratings yet — be the first to rate this tool.