Skip to content

Dify with Keln

Use Keln models in Dify apps and workflows with the Keln plugin from Dify Marketplace.

What you need to start

  • A Dify workspace, on Dify Cloud or self-hosted. A self-hosted instance needs outbound HTTPS access to api.keln.ai.
  • A Keln API key. Sign up at app.keln.ai, then open Keys → Create key.
  • Credit on your Keln account. Billing → Add credit.

Connect Dify to Keln

1. Install the Keln plugin

In Dify, open Integrations → Model Provider and type Keln in the search box. Hover over the Keln card under Install model providers and click Install, then click Install again to confirm.

The install can take a few minutes. Wait until Keln appears in your list of providers.

2. Add your Keln key

On the Keln card, choose Add API Key, paste the key and save.

Dify checks the key with one short chat request to zai-org/glm-5.3. Keln bills this request at its normal rate. If the check returns 402, your Keln balance is empty: add credit on the Billing page, then save again.

3. Pick a Keln model

Choose a Keln model in any LLM node, agent or chatbot.

Models

The plugin includes these 5 models. To use any other chat model from the Keln catalog, add it under Add another model below:

Model Input Reasoning effort
zai-org/glm-5.3 text low, high, max
zai-org/glm-5.3-flash text, images low, high, max
moonshotai/kimi-k3 text, images none, low, high, max
qwen/qwen3.8-2.4t-a95b text low, medium, xhigh
deepseek-ai/deepseek-v4.1-flash text, images none, low, high, max

Add another model

To use another chat model from the Keln catalog:

  1. On the Keln card, choose Add Model.
  2. Enter the model id from GET /v1/models as Model ID and choose LLM as Model Type.
  3. Paste your Keln key into API Key. Dify asks for it again in this form.
  4. Enter the model's context_length as Context size.
  5. Set Tool calling and Vision (image input) for the model.
  6. Click Add.

Cost

The plugin does not provide prices. Dify logs show a cost of 0 for Keln models. Keln bills your account at the rates on Models & pricing.

Usage

Your Keln Usage page shows your spend, tokens, latency and error rate. Keln stores metadata only (ZDR).

The Keln Usage page with sample data: spend, requests, tokens, time to first token, error rate and a spend-by-day chart

Where to go next

Try Keln

Create a Keln account and API key to run open models in Dify.

Get an API key Models & pricing ↗