> ## Documentation Index
> Fetch the complete documentation index at: https://docs.pullfrog.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Kimi Code

If you pay for a Kimi membership, Kimi Code lets Pullfrog run Kimi models against that membership's quota instead of a separate Moonshot API balance. Mint a key in the Kimi Code console, store it as `KIMI_API_KEY`, and pick a Kimi Code model.

```yaml theme={null}
      - name: Run agent
        uses: pullfrog/pullfrog@v0
        with:
          prompt: ${{ inputs.prompt }}
          model: kimi-for-coding/kimi-k3
        env:
          KIMI_API_KEY: ${{ secrets.KIMI_API_KEY }}
```

<Note>
  Kimi Code and the [Moonshot open platform](https://platform.moonshot.ai) are separate systems with separate keys, endpoints and billing. A `MOONSHOT_API_KEY` will not work here, and a Kimi Code key will not work on `moonshotai/*` models.
</Note>

## Setup

1. Open the [Kimi Code console](https://www.kimi.com/code/console) and create an API key. It is shown once, and you can hold at most five.
2. Store it as `KIMI_API_KEY` — on the Kimi Code card in your [Pullfrog console](https://pullfrog.com/console), or as a GitHub Actions secret mapped into the workflow's `env:` block.
3. Pick a Kimi Code model in the console, or set the `model` input as above.

## Models

Which models a key can reach depends on the membership tier. Requesting one your plan does not include returns an error naming the plan you need.

| Model             | Context                         | Plan                 |
| ----------------- | ------------------------------- | -------------------- |
| Kimi K3 256K      | 256K                            | Moderato and above   |
| Kimi K3           | up to 1M, 256K below Allegretto | Moderato and above   |
| Kimi K2           | 256K                            | every tier           |
| Kimi K2 HighSpeed | 256K                            | Allegretto and above |

Kimi K3 256K is the default Pullfrog recommends. It is the same model as Kimi K3 inside a 256K window at roughly half the quota per run, and below Allegretto that window is all Kimi K3 gets anyway — so on a Moderato plan the two return the same answers and one of them costs twice as much.

Kimi K3 accepts a [reasoning effort](/flags) of low, high or max, defaulting to high. The K2 models have no effort control and the setting is a no-op for them.

## Billing

A stored `KIMI_API_KEY` counts as bring-your-own-key: runs on Kimi Code models draw on your membership rather than being billed through [Pullfrog Router](/billing#pullfrog-router).

Router cannot serve these models — a membership is not something a router can resell — so a Kimi Code pick with no key stored falls back to the Router's default model rather than failing. The job log names the fallback.

## Troubleshooting

**Your Kimi membership doesn't include the model this run asked for.** The plan tier is too low for the model or the context window. Upgrade at [kimi.com/membership](https://www.kimi.com/membership/pricing), or switch the repo to Kimi K2, which every tier includes.

**Invalid Authentication.** Usually a Moonshot open-platform key in `KIMI_API_KEY`. Kimi Code keys come from the [Kimi Code console](https://www.kimi.com/code/console) only.

**Usage limit reached.** Kimi Code quota is windowed — five-hour, weekly and monthly caps, plus a concurrency cap. Re-trigger after the reset, or store a `MOONSHOT_API_KEY` and pick a `moonshotai/*` model to keep running on pay-as-you-go in the meantime.

## See also

* [Models](/models) — every supported provider and its key.
* [BYOK](/keys) — bringing your own provider keys.
* [Pullfrog Router](/billing#pullfrog-router) — the no-keys alternative, billed at raw provider cost.