Use MyTabulon with Kilo Code
Add MyTabulon as a custom OpenAI-compatible provider in Kilo Code, auto-fetch the models included in your tier, and start coding with one live platform key.
Fastest setup in the UI
In Kilo Code, open Settings, choose Providers, scroll to Custom provider, and enter: Provider ID mytabulon; Display name MyTabulon Coding Plan; Provider API OpenAI Compatible; Base URL https://api.mytabulon.com/v1; API key your live mtb_live_ key. Kilo fetches the available model list from GET /models. Select a model, submit, and begin a new task.
Config-file setup
Use the environment reference instead of writing the key into kilo.jsonc. The Atlas example declares its real 262K context so Kilo can compact at the right time. Add other tier models from the model catalog with their published native context values.
{
"$schema": "https://app.kilo.ai/config.json",
"model": "mytabulon/maximo-atlas-preview",
"provider": {
"mytabulon": {
"options": {
"apiKey": "{env:MYTABULON_API_KEY}",
"baseURL": "https://api.mytabulon.com/v1",
"timeout": 600000
},
"models": {
"maximo-atlas-preview": {
"name": "Maximo Atlas Preview",
"tool_call": true,
"reasoning": true,
"attachment": true,
"limit": {
"context": 262000,
"output": 65536
}
}
}
}
}
}Verify before a long task
Run kilo models and confirm that mytabulon/maximo-atlas-preview appears. If the UI cannot fetch models, verify the base URL includes /v1, the key begins mtb_live_, and the key has ai.coding. Kilo uses the Chat Completions surface, so tools, streaming, image inputs, and file inputs pass through without MyTabulon adding instructions.
export MYTABULON_API_KEY="mtb_live_..."
kilo models