Tools
The Chargeflow developer tools - OpenAPI specs, client libraries, a Postman path, the hosted AI skill for agents, plus the interactive and testing tools that are on the way.
I am looking for a Chargeflow tool, SDK, or utility. Everything below helps you build against the Chargeflow API: the specs and client tooling to generate code, the agent tools that let a coding assistant integrate for you, and the interactive and testing tools that are planned. All of it targets the API at https://api.chargeflow.io, authenticated with your API key in the x-api-key header.
| Tool | What it is | Status |
|---|---|---|
| OpenAPI specs | Machine-readable specs for the Merchants and Platforms APIs | Available |
| SDKs | Official client libraries | Coming soon |
| Postman collection | Ready-made collection of every endpoint | Coming soon (import the spec today) |
| AI skill | Hosted SKILL.md that teaches a coding agent to integrate | Available |
| MCP server | Lets an agent call Chargeflow tools directly | Coming soon |
| API explorer | In-page runner for real sandbox calls | Coming soon |
| Integration linter | Validates auth, webhook, and idempotency setup before go-live | Coming soon |
| Seeded scenarios | Ready-made sandbox test data for each dispute outcome | Coming soon |
Specs & client libraries
OpenAPI specs
Chargeflow publishes OpenAPI specifications for both APIs. Use them to generate a typed client with an OpenAPI generator, import into Postman or other tools, or review request and response schemas without making live calls. All specs describe the API served at https://api.chargeflow.io; authenticate with your API key in the x-api-key header.
Latest versioned specs (2025-04-01):
Earlier version (2024-03-18), still available:
SDKs
Official Chargeflow client libraries. None have shipped yet - Node.js / TypeScript, Python, PHP, and Ruby are planned, and package names and install instructions will be listed here when each one ships. In the meantime the Chargeflow API is a standard REST API that works with any HTTP client, so you can integrate today: use the quickstart for copy-paste curl examples, or generate a typed client yourself from the OpenAPI specs above.
Postman collection
An official Postman collection covering every Chargeflow endpoint is in progress; a download link will appear here when it ships. Until then, import an OpenAPI spec into Postman to get a request per endpoint: choose Import, paste the spec URL or upload the file, and Postman builds a collection from it. After importing, set your x-api-key header so requests are authenticated - see Test credentials for sandbox keys.
Agent tools
AI skill
A hosted, self-contained reference file that teaches a coding agent to integrate Chargeflow - authentication, listing and enriching disputes, uploading evidence, handling pre-chargeback alerts, and webhooks - routing each intent to the exact endpoint instead of guessing. It is hosted at /SKILL.md: point your agent at that URL, or paste its contents into your assistant's context. You can also copy the ready-made prompt from the Setup for agents button on the docs homepage.
The skill covers the base URL, auth, and versioning (the x-api-key header and the credential health check to run first), merchant and platform intent-to-endpoint tables, webhooks, and the rules an agent must follow (never invent endpoints, honor Retry-After on 429s, keep the key server-side, treat the OpenAPI specs as source of truth). Every docs page is also fetchable as plain markdown - append .md to any URL - and the skill links to /llms.txt as the full docs index.
MCP server
An MCP server is coming so a coding agent can call Chargeflow tools directly from Claude Code, Cursor, Codex, and VS Code. When it ships, this section becomes a per-client setup selector with the exact install command or JSON config for each client. Until then, the agent path that already works needs no MCP: give your agent the AI skill at /SKILL.md, point it at /llms.txt as the docs index, and hand it the OpenAPI specs for exact request and response shapes.
Interactive & testing tools
These tools are planned; each entry lists what already works in the meantime.
API explorer
An in-page explorer that runs real calls against the sandbox, with your test keys injected automatically once you sign in. Until it ships, use the quickstart to make your first call and test credentials to authenticate.
Integration linter
A check that validates your authentication, webhook, and idempotency setup and flags anything that would fail in production. Until it ships, work through the go-live checklist to confirm your integration is ready.
Seeded scenarios
Ready-made sandbox test scenarios that let you exercise each dispute outcome without waiting for real events - planned coverage includes a dispute won, a dispute lost, a representment, an alert that prevents a chargeback, and an EaaS package. Until they ship, generate a test key and call the sandbox directly to set up your own test data.