Run Kortyx Studio Locally
Updated 17 minutes ago • August 2, 2026
This is the shortest path from a working Kortyx SDK application to its first observable run.
By the end, you will have:
- Studio, the telemetry API, and PostgreSQL running through Docker;
- one server-side Kortyx agent connected; and
- a run visible at
http://localhost:6300.
Before you start
You need:
- Node.js 22 or newer;
- npm 10 or newer;
- Docker Engine with Docker Compose v2, or a current Docker Desktop; and
- free local ports
6300for Studio and6400for the telemetry API.
Linux containers on AMD64 and ARM64 are supported. Docker Desktop on Apple Silicon selects the ARM64 images automatically.
1. Start Studio
Run this from your Kortyx SDK project:
The first start can take a few minutes while Docker downloads the images. The command then migrates and bootstraps PostgreSQL, starts the services, waits for health checks, and prints the connection details.
The exact secret values are unique to your installation.
Keep them private: Local credentials are stored under
~/.kortyx/studio. Do not commit the generated.envfile or expose a telemetry key through a browser bundle.
2. Connect your application
Install the Studio telemetry adapter:
Copy the four KORTYX_TELEMETRY_* values printed by the CLI into the server-only environment of your application. Then create the telemetry adapter once and attach it to your agent.
For a fuller explanation of identifiers and content capture, see Connect Your Project.
3. See your first run
- Restart your application so it reads the new environment values.
- Trigger any request that executes the Kortyx agent.
- Open http://localhost:6300.
- Sign in with the username and password printed by the CLI.
- Open Runs and select the new run.
You are connected when the run appears. Studio can now explain the workflow path, model calls, timing, token usage, interrupts, and captured payloads available for that execution.
If the run does not appear
Check the installation before changing application code:
Then verify:
- the API URL and key are available to the server process that creates the agent;
- the application was restarted after its environment changed;
- the telemetry API is reachable from that server process; and
- the Runs time filter includes the current time.
An Invalid telemetry API key error means the application or Studio is presenting a key that does not match the verifier in this installation. Print the current values with npx kortyx studio credentials; if you intentionally rotate them, update every affected server process.
Next steps
- Learn the local lifecycle commands in CLI Commands.
- Decide whether payload content may be captured in Connect Your Project.
- Read Operations and Troubleshooting before backup, reset, or upgrade.
- Use Deploy on a Server when Studio must be reachable beyond your machine.