LLM providers offer the models used by chats, agents, and AI-assisted workflows.
Current support on this page reflects the provider definitions and curated model catalog shipped with Kontext.
Provider Support
| Provider | Type | Required Info | Custom Endpoint |
|---|
| LM Studio | Local | endpointBaseUrl | Yes |
| Ollama | Local | endpointBaseUrl | Yes |
| OpenAI | Hosted | apiKeyRef | Yes |
| Anthropic | Hosted | apiKeyRef | No |
| Google Gemini | Hosted | apiKeyRef | No |
| AWS Bedrock | Hosted | region, credentials | No |
| OpenRouter | Hosted | apiKeyRef, endpointBaseUrl | Yes |
| OpenAI-Compatible | Hosted | preset, apiKeyRef, endpointBaseUrl | Yes |
Curated Models
Local providers are first-class. Current curated local/open-weight examples include openai/gpt-oss-20b and google/gemma-4-26b-a4b.
| Provider path | Example curated models |
|---|
lm_studio, ollama | openai/gpt-oss-20b, google/gemma-4-26b-a4b |
openai | openai/gpt-5.4-mini, openai/gpt-4o-mini |
anthropic | anthropic/claude-sonnet-4-6, anthropic/claude-haiku-4-5-20251001, anthropic/claude-3-5-sonnet-latest |
gemini | gemini/gemini-2.5-flash |
bedrock | bedrock/amazon.nova-2-lite-v1:0 |
openrouter | openai/gpt-4o-mini |
openai_compatible:deepseek | deepseek/deepseek-chat |
openai_compatible:mistral | mistral/mistral-small-latest |
OpenAI-Compatible Presets
| Preset | Default endpoint |
|---|
deepseek | https://api.deepseek.com/v1 |
mistral | https://api.mistral.ai/v1 |
Config Fields
| Field | Purpose |
|---|
endpointBaseUrl | Base URL for local or compatible providers. |
providerModelName | Provider-specific model identifier when it differs from the curated catalog model ID. |
region | Required for Bedrock connections. |
compatibilityPresetId | Required for openai_compatible connections using a curated preset. |
defaultSampling | Default temperature, topP, maxTokens, and reasoning settings. |
Available In
| Surface | Usage |
|---|
| Chat sessions | Default provider for interactive chat and agent work |
| Data Agent | Provider for team and plan execution |
| AI-assisted workflows | Shared provider selection for AI features |
