Every week I watch developers paste production secrets into random websites to decode a JWT, estimate token costs by doing math in a spreadsheet, and pay $20/month for a prompt testing tool that could be 50 lines of JavaScript running locally in a browser tab.
It bothered me enough that I built something about it.
Over the past few months I shipped 71 free browser-based tools for anyone working with AI models, cloud infrastructure, or code. Every tool runs 100% client-side. No backend. No sign-up. No tracking. Your data never leaves your machine.
Here's what's in the toolkit and how to actually use each one.
AI and LLM Tools (31 tools)
These are the tools I use daily working with Claude, GPT, Gemini, and DeepSeek across production infrastructure.
1. Token Counter
Paste any text and get an instant token estimate across 23 LLM models with cost calculations. I use this before every API call to make sure I'm not blowing past context limits or burning money on a model that's overkill for the task.
Real use case: Before sending a 12,000-word document to Claude 4.6 Opus for analysis, I check the token count and cost first. Turns out the same job costs $0.03 on Gemini 2.5 Flash.
2. AI Model Comparison
A sortable, filterable table of 23 LLM models with context windows, pricing per million tokens, capabilities, and release dates. Side-by-side comparison without opening 6 different pricing pages.
Real use case: Our team was debating GPT-4.1 vs Claude 4 Sonnet for code review automation. One look at this table settled the argument in 30 seconds.
3. AI Pricing Calculator
Set your average requests per day, input tokens, and output tokens. It calculates daily, monthly, and yearly costs across 19 models from OpenAI, Anthropic, Google, DeepSeek, Meta, and Groq.
Real use case: My manager asked for a cost projection on an AI-powered incident triage system. I plugged in our numbers and had a comparison spreadsheet in under a minute.
4. Context Window Visualizer
Paste your prompt and see a visual bar chart showing how much of each model's context window it fills. Covers 10 models including GPT-5.3 Codex, Claude 4.6 Opus, Gemini 2.5 Pro, and Llama 4 Maverick.
Real use case: I was feeding a full Terraform module (14,000 tokens) plus system prompt (800 tokens) into DeepSeek R1 and getting truncated output. The visualizer showed I was at 94% capacity. Switching to Claude 4 Sonnet gave me 85% headroom.
5. System Prompt Builder
A visual builder with togglable sections: role, context, task, constraints, output format, tone, examples, and safety rules. Five pre-built persona templates. Live word and token count.
Real use case: I built a system prompt for an AI agent that reviews pull requests for Terraform security misconfigurations. The builder caught that I forgot to add output format constraints, which was causing inconsistent responses.
6. Prompt A/B Comparator
Write two versions of a prompt side by side. Get clarity scores, specificity scores, token counts, cost estimates, and a checklist of prompt engineering best practices (role defined? output format specified? examples included?).
Real use case: I had a vague one-liner prompt that kept producing generic output. Rewrote it with constraints and examples. The comparator confirmed the new version scored 85% vs 15% on specificity. The output quality difference was night and day.
7. LLM Parameter Guide
An interactive explainer for temperature, top_p, top_k, frequency penalty, presence penalty, and max tokens. Drag sliders and see example output at different values. Shows which providers support which parameters.
Real use case: A junior engineer on my team kept getting wildly different outputs from the same prompt. He had temperature set to 1.8. Five minutes with this guide and he understood why 0.2 was the right setting for code generation.
8. Prompt Injection Scanner
Paste any prompt or user input and scan it against 18 known injection patterns: instruction overrides, DAN/jailbreak attempts, control token injection, role hijacking, encoded payloads, and JSON structure escapes. Returns a risk score from 0-100.
Real use case: We're building an internal chatbot that takes user input. Before shipping, I ran our test inputs through the scanner. It caught a subtle control token injection pattern (<|im_start|>system) that our regex filter missed.
9. AI Output Parser
Paste raw AI output and extract structured data. Auto-detects and handles: JSON buried in markdown code blocks, markdown tables to CSV, code block extraction, <thinking> tag stripping, and numbered list to JSON array conversion.
Real use case: Claude returns JSON wrapped in ```json fences with explanation text before and after. Instead of writing regex every time, I paste the raw output and get clean JSON in one click.
10. Fine-tuning Data Formatter
Convert CSV or manually entered training examples to JSONL format for fine-tuning. Supports OpenAI (chat format), Anthropic (messages format), and Together AI (Llama instruct format). Token estimation included.
Real use case: I had 200 rows of Q&A training data in a Google Sheet. Exported to CSV, pasted it in, selected "OpenAI", and had a valid JSONL file ready to upload in seconds.
11. AI API Request Builder
Pick a provider (OpenAI, Anthropic, Google Gemini, DeepSeek, Groq), configure model, system prompt, temperature, max tokens, and streaming. Export as cURL, Python, JavaScript, or raw JSON.
Real use case: I needed to quickly test a Groq API call with Llama and couldn't remember their exact request format. Selected Groq, configured the params, copied the cURL, and ran it from my terminal.
12. JSON Schema for AI
A visual editor for building JSON schemas used in OpenAI function calling and structured output. Define fields, types, descriptions, enum values, and required flags. Export as OpenAI function tool, structured output, or standard JSON Schema. Five starter templates.
13. AI Content Detector
Heuristic analysis that scores text for AI-generated patterns: sentence length uniformity, AI buzzword density, passive voice percentage, vocabulary richness, repetitive sentence starters, and hedging language frequency.
14. Image Prompt Builder
Structured prompt builder for Midjourney v7, DALL-E 3, and Stable Diffusion. Dropdown selectors for subject, style, lighting, camera angle, mood, color palette, and quality modifiers. Outputs platform-specific syntax.
15. Prompt Templates
10 copy-paste prompt templates for DevOps and infrastructure workflows: Terraform security review, Kubernetes troubleshooting, CI/CD pipeline design, incident post-mortem analysis, architecture review, and more.
16. README Generator
Pick a project template (Node.js API, Python Package, React App, CLI Tool, Terraform Module, Docker Image). Toggle 15 sections. Auto-generates badges, install commands, and structure based on your stack.
17. AI Resume Builder
Paste a job description and your experience. The tool extracts keywords from the job posting, matches them against your background, and generates a tailored resume with matched skills highlighted. Four output templates: Technical, Executive, Creative, and ATS-Optimized.
Real use case: I pasted a Senior DevOps Engineer job posting and my experience. The tool extracted 23 keywords (Kubernetes, Terraform, CI/CD, etc.), matched 18 of them against my experience, and generated an ATS-friendly resume that emphasized the right skills. 78% keyword match score.
18. Meeting Notes Summarizer
Paste raw meeting notes or a transcript. The tool uses heuristic pattern matching to extract attendees (from @mentions and "Name:" patterns), key decisions ("decided", "agreed", "confirmed"), action items ("will", "should", "need to" with owner attribution), and follow-ups.
Real use case: After a 45-minute sprint planning call, I pasted the raw notes. It pulled out 4 attendees, 3 decisions, 7 action items with owners, and 2 follow-ups. Saved me 10 minutes of manual formatting.
19. AI Tone Rewriter
Rewrite any text in six tones: formal, casual, friendly, assertive, diplomatic, and technical. Uses client-side word and phrase replacement maps. Shows change count and word count delta.
Real use case: I had a direct feedback message that came across as harsh. Ran it through the diplomatic tone and it softened "you're wrong about the deadline" to "I see it differently regarding the timeline." Same message, zero conflict.
20. i18n Scaffold Generator
Paste your English i18n JSON keys and generate translation file scaffolds for 15+ languages across four framework formats: React-Intl, next-intl, i18next, and Vue-i18n. Detects placeholders ({name}, {{count}}) and preserves them. Compare existing locale files to find missing keys.
Real use case: We had 45 translation keys in English for a Next.js app. I pasted the JSON, selected 6 target languages and next-intl format, and had scaffold files for all of them ready for the translation team in seconds.
21. AI Data Extractor
Paste any unstructured text and extract structured entities: emails, URLs, phone numbers, dates (ISO, US, European formats), IP addresses, currency amounts, and hex colors. Export as JSON or CSV.
Real use case: I pasted a raw email thread with 200 lines and extracted 12 email addresses, 8 URLs, 3 phone numbers, and 15 dates. Exported to CSV for a CRM import. Would have taken 20 minutes manually.
22. AI Code Reviewer
Paste a code snippet, select the language (JS/TS, Python, Go, Java), and get a static analysis review. Detects hardcoded secrets, SQL injection patterns, eval()/innerHTML usage, console.log statements, deep nesting, magic numbers, long functions, missing error handling, and TODO/FIXME comments. Returns severity levels with line numbers.
Real use case: Pasted a 150-line Express handler before a PR review. It caught a hardcoded API key on line 23, a SQL injection pattern on line 67, and 4 console.log statements I forgot to remove. Overall score: 45/100.
23. Git Commit & PR Writer
Paste a git diff or describe your changes. The tool parses file names, change types, and auto-detects the conventional commit type (feat/fix/refactor/docs/chore). Generates a conventional commit message and a full PR description with Summary, Changes, and Testing sections.
Real use case: Pasted a 300-line diff from a feature branch. It detected 8 files changed, auto-classified it as feat, and generated a commit message with scope and a GitHub PR description I could copy directly.
24. AI Regex Builder
Describe what you want to match in plain English (like "email addresses" or "US phone numbers") and get the regex pattern with a detailed explanation of each part. Includes a library of 30+ pre-built patterns and a live test area with match highlighting.
Real use case: Needed a regex for matching Semantic Version strings in CI output. Typed "semver" and instantly got /^(0|[1-9]\d*)\.(0|[1-9]\d*)\.(0|[1-9]\d*)(?:-((?:0|[1-9]\d*|\d*[a-zA-Z-]...)...))?$/ with each capture group explained.
25. API Docs Generator
Paste Express routes, FastAPI endpoints, Go handlers, or TypeScript interfaces. The tool parses your code to extract endpoints, methods, parameters, and types, then generates structured Markdown documentation or OpenAPI 3.0 YAML.
Real use case: Pasted 5 Express route handlers and got complete Markdown docs with endpoint, method, parameters table, request/response examples, error codes, and curl examples for each route.
26. AI SQL Assistant
Works in two directions: describe a query in English and get SQL, or paste SQL and get an English explanation. Supports PostgreSQL, MySQL, and SQLite dialects. Handles SELECT, JOIN, GROUP BY, window functions (ROW_NUMBER, RANK, LAG, LEAD), CTEs, subqueries, and more.
Real use case: Typed "find all users who signed up in the last 30 days, grouped by country, showing the count, ordered by count descending, limit 10" and got a correct PostgreSQL query with proper date arithmetic.
27. Document Q&A
Paste a long document and ask questions about it. The tool uses TF-IDF keyword matching to find and rank the most relevant paragraphs. Returns the top 5 passages with relevance scores, not hallucinated answers.
Real use case: Pasted a 40-page AWS Well-Architected Framework excerpt and asked "what are the reliability best practices?" Got 5 relevant paragraphs ranked by relevance, each with a percentage score and paragraph number.
28. Automation Recipe Builder
Describe an automation in English or browse a library of 25 pre-built recipes across 5 platforms: Cron jobs, GitHub Actions, AWS Lambda, Shell scripts, and Terraform. Every recipe is complete, runnable code with inline comments.
Real use case: Needed a GitHub Action that runs daily to check for outdated npm dependencies. Selected the "Dependency Update" recipe from the library and had a complete workflow YAML with PR creation, labels, and a cron schedule.
29. AI Log Analyzer
Paste log lines and get an instant analysis: error count by severity, top errors by frequency, timeline visualization, stack trace extraction, and root cause suggestions. Auto-detects log format (JSON, syslog, Apache, nginx) and parses accordingly.
Real use case: Pasted 500 lines from a failing Node.js service. The analyzer found 23 errors (12 unique), detected an OOM pattern, and suggested checking memory limits and looking for memory leaks. The timeline showed errors spiking at 3:00 AM, which matched a cron job.
30. Prompt Eval Suite
Score your prompts against 12 engineering best practices: role definition, context, output format, examples, constraints, tone guidance, specificity, length, error handling, injection safety, and evaluation criteria. Get a 0-100 score with detailed improvement suggestions.
Real use case: Scored a production system prompt and got 42/100. It was missing output format constraints, examples, and error handling instructions. After fixing those three things, the score jumped to 85 and the model's output consistency improved dramatically.
31. AI Diagram Generator
Select a diagram type (flowchart, sequence, ER, class, state, Gantt), pick from 8 pre-built templates (CI/CD pipeline, microservices architecture, OAuth flow, database schema, git branching), or write Mermaid code directly. Live rendered SVG preview with copy and export.
Real use case: Needed a sequence diagram for an OAuth flow to include in a design doc. Selected the template, tweaked a few labels, and had a clean Mermaid diagram rendered in the browser. Copied the code and pasted it into the GitHub PR description.
SEO and Web Tools (4 tools)
32. Meta Tag Generator
Generate SEO meta tags, Open Graph tags, and Twitter Card tags for any webpage. Real-time character count feedback for title (60 char limit) and description (160 char limit). Copy the HTML and paste it into your <head>.
33. Social Card Preview
See exactly how your page will look when shared on X/Twitter, Facebook, LinkedIn, and Slack. Paste your OG data and switch between platform mockups.
Real use case: I previewed my blog post's social card before publishing and caught that my description was getting truncated on Twitter. Fixed it before a single share went out.
34. Robots.txt Generator
Visual builder with presets: Allow All, Block All, Standard (block /api and /admin), Block AI Crawlers (GPTBot, ChatGPT-User, CCBot, anthropic-ai, Google-Extended), and E-commerce. Add custom rules per user-agent.
35. Color Contrast Checker
Pick foreground and background colors, get the exact contrast ratio, and see pass/fail indicators for WCAG 2.1 AA and AAA standards. Live text preview at 24px, 16px, and 12px.
Data Converters (7 tools)
36. JSON/YAML Converter
Paste JSON, get YAML. Paste YAML, get JSON. I use this multiple times a day converting between Kubernetes manifests and API payloads.
37. JSON Formatter
Format, minify, validate, and alphabetically sort JSON. Shows key count and nesting depth.
38. YAML Validator
Validates YAML syntax and reports the exact line with the error. Essential for catching that one missing indent in a 300-line Kubernetes deployment.
39. JSON to TypeScript
Paste any JSON object and get TypeScript interfaces generated instantly. Handles nested objects, arrays, nullable types, and optional export and type vs interface output.
40. CSV to JSON (and back)
Bidirectional converter with support for comma, tab, semicolon, and pipe delimiters. Handles quoted fields and special characters.
41. Base64 Encoder/Decoder
Encode and decode Base64 with proper UTF-8 handling. I use this weekly for Kubernetes secrets.
42. URL Encoder/Decoder
Encode and decode URL components. Essential for debugging API query parameters.
DevOps and CI/CD Tools (7 tools)
43. Cron Expression Builder
A visual builder where you click fields instead of memorizing cron syntax. Shows the next 10 execution times and includes common presets (every hour, daily at midnight, every Monday at 9am).
44. Docker Run to Compose
Paste a docker run command and get a valid docker-compose.yml. Handles volumes, ports, environment variables, networks, and restart policies.
Real use case: A teammate shared a 200-character docker run command in Slack. I pasted it into this tool and had a clean compose file committed to the repo in 30 seconds.
45. Subnet Calculator
Enter a CIDR block and get the network address, broadcast address, usable host range, total hosts, wildcard mask, and subnet mask. Every cloud engineer needs this.
46. Chmod Calculator
Toggle permission bits visually or type a numeric value (supports both 3-digit and 4-digit like 0755). See the symbolic notation update in real time.
47. HTTP Status Code Reference
Every HTTP status code with descriptions, searchable and filterable. When your API returns a 422 and you can't remember what it means, this is faster than Googling it.
48. Diff Checker
Side-by-side and unified diff views with line numbers. I use this for comparing Terraform plan outputs, config file versions, and PR changes.
49. Regex Tester
Write a pattern, paste test text, see matches highlighted live with capture group breakdown. Supports flags and shows match count.
Security Tools (4 tools)
50. JWT Decoder
Paste a JWT and instantly see the decoded header, payload, algorithm, and expiration status. No third-party service touches your token.
Why this matters: Every JWT decoder website you've used before sends your token to their server. This one doesn't. Your auth tokens stay in your browser.
51. Hash Generator
Generate SHA-256, SHA-384, and SHA-512 hashes using the browser's native Web Crypto API. Useful for verifying file integrity or generating content hashes.
52. UUID/ULID Generator
Generate UUID v4, ULID (sortable), or NanoID with custom length. Bulk generation supported.
53. Password Generator
Create cryptographically secure passwords with configurable length, character types, and a visual strength meter.
Text and Data Tools (3 tools)
54. Unix Timestamp Converter
Convert Unix timestamps to human-readable dates and back. Shows a live clock with the current timestamp. I keep this open in a tab whenever I'm debugging CloudWatch logs.
55. String Utils
Convert between camelCase, snake_case, kebab-case, PascalCase, UPPER_CASE, and more. Sort lines, remove duplicates, count characters and words.
56. Markdown Preview
Live Markdown editor with a rendered preview. Useful for drafting README files, documentation, and blog posts before committing.
The Stack Behind It
These tools are built with Next.js and React. The heavy lifting happens through native browser APIs:
- Web Crypto API for hashing and secure random generation
- TextEncoder/TextDecoder for Base64 and encoding
- RegExp for pattern matching
- Intl.DateTimeFormat for timestamp conversion
- CSS color parsing for contrast calculations
A handful of lightweight open-source libraries handle the edge cases: js-yaml for YAML parsing, cronstrue for human-readable cron descriptions, diff for the diff checker, and mermaid for diagram rendering. Everything else is vanilla JavaScript.
No server. No database. No cookies. No analytics on your tool usage.
Why I Built These
I got tired of three things:
-
Pasting sensitive data into random websites. JWT tokens, API keys in Base64, production config files. Every "free online tool" is a server you don't control. These tools never send your data anywhere because there is no server.
-
Paying for basic utilities. Token counters, prompt testers, JSON formatters. These are simple tools that don't need a SaaS subscription, a login wall, or a "free tier" with limits.
-
Switching between 15 different sites. Every tool here is one click away from every other tool. No new tabs, no new accounts, no new cookie banners.
If you're an engineer working with AI models, cloud infrastructure, or code, bookmark the toolkit and stop giving your data to websites you don't need to trust.
All 71 tools are free and will stay free. If they save you time, share this post with your team or buy me a coffee.