A
AllKit
Built for AI agents

The API that gives your AI superpowers

23 REST endpoints your LLM can call right now. Hash data, decode JWTs, remove backgrounds, transcribe audio, generate images — all with a single HTTP request. No SDK, no signup, no API key required for free tier.

API DocumentationView /llms.txtPricing — $9/mo unlimited

Why AI agents need a tool API

LLMs can't process media

Your AI can't hash a string, decode a JWT, or remove an image background natively. AllKit gives it those capabilities via simple API calls.

Zero infrastructure

No servers to deploy, no GPU to rent, no models to host. Just HTTP POST with JSON. Works from any language, any framework, any platform.

Discoverable by design

Our /llms.txt file lets AI agents discover all 37 tools automatically. No manual configuration needed.

Works with every AI platform

Claude Code & MCP

Claude Code can call AllKit endpoints via tool-use. Hash a file, decode a JWT, remove a background — all inside your coding session.

curl -X POST https://allkit.dev/api/v1/hash \
  -H "Content-Type: application/json" \
  -d '{"text": "verify this content"}'

ChatGPT & Custom GPTs

Build a Custom GPT that calls AllKit to process images, transcribe audio, or generate QR codes for users — no backend needed.

// OpenAI function calling schema
{
  "name": "remove_background",
  "url": "https://allkit.dev/api/v1/ai/remove-background",
  "method": "POST",
  "body": { "image": "<base64>" }
}

AI Agent Frameworks

LangChain, CrewAI, AutoGen, Vercel AI SDK — any agent framework that supports HTTP tools can use AllKit in seconds.

# Python — any HTTP client works
import requests

r = requests.post(
    "https://allkit.dev/api/v1/ai/speech-to-text",
    json={"audio": base64_audio, "task": "transcribe"}
)
print(r.json()["result"])

Perplexity, Copilot & Others

Any LLM that discovers tools via /llms.txt can find AllKit automatically and call endpoints on behalf of users.

# AllKit is discoverable at:
https://allkit.dev/llms.txt

# Machine-readable inventory of all 37 tools
# with exact API calls and parameters

23 endpoints, one base URL

https://allkit.dev/api/v1

Developer Tools

POST/api/v1/hash
POST/api/v1/base64
POST/api/v1/json-format
GET/api/v1/uuid
POST/api/v1/jwt-decode
POST/api/v1/url-encode
POST/api/v1/timestamp
POST/api/v1/csv-json
GET/api/v1/password
POST/api/v1/word-count

AI Media & Processing

POST/api/v1/ai/remove-background
POST/api/v1/ai/ocr
POST/api/v1/ai/generate-image
POST/api/v1/ai/upscale
POST/api/v1/ai/face-swap
POST/api/v1/ai/tts
POST/api/v1/ai/speech-to-text
POST/api/v1/ai/voice-clone
POST/api/v1/ai/live-portrait
POST/api/v1/ai/image-to-video

AI Smart Tools

POST/api/v1/ai/regex
POST/api/v1/ai/cron
POST/api/v1/ai/privacy-policy

Start free, scale when ready

3 API requests per day — free, no signup, no API key. When your agent needs more, upgrade to Pro for unlimited access at $9/month.

Get unlimited API access — $9/moRead the docs

What is an AI Tool-Use API?

AI tool-use (also called function calling) is how modern LLMs interact with the real world. When ChatGPT, Claude, or any AI assistant needs to perform a task it can't do natively — like generating a hash, removing an image background, or transcribing audio — it calls an external API. AllKit provides that API.

Unlike traditional APIs that are designed for human developers to integrate into applications, AllKit's API is specifically designed for machine consumption. Every endpoint follows consistent patterns: JSON in, JSON out. Every response includes clear error messages that an LLM can interpret and communicate to the user. And our /llms.txt file provides a machine-readable inventory of every tool and endpoint, so AI agents can discover capabilities automatically.

How LLMs Discover and Use AllKit

The /llms.txtstandard is an emerging convention (similar to robots.txt) that tells AI agents what a website can do for them. When an LLM encounters a task it needs help with, it can check /llms.txt to find relevant tools. AllKit's llms.txt lists all 37 tools with exact API endpoints, request formats, and descriptions — everything an AI needs to make the right call.

This is particularly powerful for AI coding assistants like Claude Code. When a developer asks "hash this string with SHA-256" or "decode this JWT", the assistant can call AllKit's API directly instead of writing code to do it. For image tasks like background removal or OCR, the value is even clearer — these require ML models that no coding assistant has built in.

Supported AI Platforms

AllKit works with any system that can make HTTP requests. This includes OpenAI's function calling (Custom GPTs, Assistants API), Anthropic's tool use (Claude, Claude Code), Google's Gemini tools, and every agent framework: LangChain, CrewAI, AutoGen, Semantic Kernel, Vercel AI SDK, and more. If your agent can call a URL, it can use AllKit.

Why AllKit vs. Building Your Own

You could deploy your own hash function, your own OCR model, your own background remover. But that means managing servers, GPU instances, model updates, scaling, and monitoring. AllKit handles all of that. One API, 23 endpoints, zero infrastructure. Your agent calls the endpoint, gets the result, moves on.

For AI media tools specifically (background removal, image generation, speech-to-text, voice cloning), the infrastructure cost of running GPU models is significant. AllKit absorbs that cost in the free tier and offers unlimited access at $9/month — a fraction of what it would cost to run these models yourself.

No Other Tool Site Offers This

We checked every major competitor: iLovePDF, TinyWow, FreeConvert, IT-Tools, SmallDev.tools, Transform.tools. None of them have a public API designed for LLM tool-use. None have /llms.txt. None have endpoints that an AI agent can call. AllKit is the only tool platform built from the ground up for the AI-native era.