Build AI assistants
that actually work
Power OpenClaw, Hermes Agent, or your own chatbot with Kyma. One API key, active models, streaming, tool use. Pay per use.
Agent-Ready Models
Qwen 3.6 Plus for strong general reasoning, Kimi K2.6 for tool use, and DeepSeek R1 for deeper planning workflows.
Streaming & Tool Use
Full streaming support, function calling, JSON mode. Everything your agent framework needs, OpenAI-compatible.
Multilingual models
Use Qwen models for multilingual tasks. Evaluate the output in the languages your audience uses. Build multilingual assistants effortlessly.
Best Models for Assistants
Ranked by conversational quality, reasoning depth, and real user adoption.
Supports reasoning, tool use, vision input. Compare current prices and measured uptime on the model page.
qwen-3.6-plusReasoning
Supports tool use. Compare current prices and measured uptime on the model page.
llama-3.3-70bTool use
Supports reasoning, tool use, vision input. Compare current prices and measured uptime on the model page.
gemma-4-31bReasoning
Supports reasoning, tool use. Compare current prices and measured uptime on the model page.
deepseek-r1Reasoning
Supports reasoning, tool use, vision input. Compare current prices and measured uptime on the model page.
gemini-3-flashReasoning
Setup guide
1. Create account → 2. Pick your tool below → 3. Paste the config.
Settings → Providers → OpenAI Compatible Base URL: https://kymaapi.com/v1 API Key: kyma-xxxxx (your Kyma key) Model: qwen-3.6-plus Tip: OpenClaw works best with qwen-3.6-plus or llama-3.3-70b for general tasks.
# .hermes/config.yaml provider: type: openai-compatible base_url: https://kymaapi.com/v1 api_key: kyma-xxxxx # your Kyma key model: qwen-3.6-plus
from openai import OpenAI
client = OpenAI(
base_url="https://kymaapi.com/v1",
api_key="kyma-xxxxx" # your Kyma key
)
# Multi-turn conversation
messages = [
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": "Summarize this document..."},
]
response = client.chat.completions.create(
model="qwen-3.6-plus",
messages=messages,
stream=True
)from langchain_openai import ChatOpenAI
llm = ChatOpenAI(
base_url="https://kymaapi.com/v1",
api_key="kyma-xxxxx", # your Kyma key
model="qwen-3.6-plus",
streaming=True
)
response = llm.invoke("Plan a 3-step workflow...")import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://kymaapi.com/v1",
apiKey: "kyma-xxxxx", // your Kyma key
});
// Streaming assistant
const stream = await client.chat.completions.create({
model: "qwen-3.6-plus",
messages: [
{ role: "system", content: "You are a helpful assistant." },
{ role: "user", content: "Help me plan..." },
],
stream: true,
});# Any OpenAI-compatible chatbot framework:
Base URL: https://kymaapi.com/v1
API Key: kyma-xxxxx
Model: qwen-3.6-plus (best quality)
or gemma-4-31b (vision input)
or llama-3.3-70b (chat)
# Supports: streaming, function calling,
# tool use, JSON mode, system promptsWhat People Build
🤖 Personal AI Assistant
Multi-agent teams with OpenClaw: planning, research, coding, all orchestrated by AI.
💬 Customer Support Bot
Multilingual chatbot that understands context, handles complex queries, and escalates when needed.
📊 Data Analysis Agent
Text-to-SQL pipeline that converts questions into database queries. Popular with MiniMax M2.5.
✍️ Content Creation
Audio drama scripts, blog posts, social media: powered by Gemini 3 Flash's 1M context window.
Build your AI assistant with Kyma API
$0.50 signup credit for eligible free-tier models after email verification. No card required.
Create accountExplore more
Dig into the models behind these assistants, compare options, and check live rankings.