JWT Decoder Skill

What this skill does

Let your agent inspect JWT tokens. Call it as an MCP tool to decode header and payload claims and check expiry — no signature verification. You switch it on with a single toggle — no SDK and no endpoint wiring — and it answers over MCP or REST using the same key as every other skill.

See it work

A real response from the JWT Decoder Skill — exactly what your agent gets back when it makes the call.

vervekit · jwtdecodertool call

{
  "token": "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzdWIiOiIxMjM0NTY3ODkwIiwibmFtZSI6IkpvaG4gRG9lIiwiaWF0IjoxNTE2MjM5MDIyfQ.SflKxwRJSMeKKF2QT4fwpMeJf36POk6yJV_adQssw5c"
}

result

header: { … }
payload: { … }
signature: SflKxwRJSMeKKF2QT4fwpMeJf36POk6yJV_adQss…
isExpired: false
expiresAt: null
issuedAt: 2018-01-18T01:30:22.000Z
tokenAge: 2557 days
warning: This API only decodes JWT tokens. It doe…

Run it with your own input. Live calls happen in your dashboard, on your key.

Ways to use it

Convert without a parser

Enable JWT Decoder so your app can decode jwt in one call — no format library to pull in, no edge cases to hand-roll.

A tool your agent can call

Over MCP, jwtdecoder lets your model convert data mid-task, on the same connection as every other skill.

Batch from your backend

Call JWT Decoder over REST to decode jwt across a whole queue in a server-side job.

Use it in your app

Once enabled, this skill is reachable two ways — pick whichever fits how you build. Both use the same key.

For AI agentsMCP

It appears to your model as a callable tool. No extra code — the agent invokes it when a task needs it.

jwtdecoder

For apps & backendsREST

Call it from your server with one request and your key. Node, Python, Go — anything that can send a GET.

GET /v1/jwtdecoder

Ask your agent

“Enable JWT Decoder and ask your agent to decode jwt.”
“Convert this payload into the format I need.”

JWT Decoder Skill, answered

How to connect it over MCP or REST.

How do I add JWT Decoder to my app or agent?

Enable the JWT Decoder Skill on VerveKit, then reach it two ways with the same key: over MCP (it appears to your agent as the jwtdecoder tool) or over REST (call it from any backend). No SDK to install and no endpoint to wire.

MCP or REST — which should I use?

Both work off one key. Use MCP when an AI agent should decide when to decode jwt — the skill shows up as a callable tool. Use REST when your own server-side code should call it directly. Many apps use both.

Which agents and frameworks does it work with?

Any MCP client — Claude, Cursor, LangChain, and custom agents all speak the Model Context Protocol, so JWT Decoder appears as a standard tool with nothing skill-specific to integrate.

How many credits does a JWT Decoder call cost?

Each call costs 1 credit. Every skill rides the same key and the same connection, so enabling more skills doesn't add more integrations to manage.

Do I need to install anything?

No SDK and no endpoint wiring — enabling the skill is a toggle. You point your agent at the VerveKit MCP endpoint (or call REST) and JWT Decoder is available immediately.

Where does the data come from, and what shows on my bill?

VerveKit runs on APIVerve, our production data engine; JWT Decoder is one of 300+ skills on the same key. Invoices and card statements show APIVERVE.

Convert without a parser

A tool your agent can call

Batch from your backend

JWT Decoder Skill, answered

Give your software a way to act on the world.

Scaling in production?