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

# AI Providers Overview

> Configure AI providers in EnConvo

## Overview

EnConvo supports a wide range of AI providers, giving you the flexibility to choose the best model for your needs. You can use cloud-based APIs or run models locally for complete privacy.

## Setup Steps

<Steps>
  <Step title="Open Settings">
    Go to **Settings** → **AI Model**
  </Step>

  <Step title="Choose a provider">
    Select a provider such as OpenAI, Anthropic, Gemini, EnConvo Cloud, OpenRouter, Ollama, or LM Studio.
  </Step>

  <Step title="Configure Credentials">
    Open the provider's credential settings and enter the required API key, base URL, OAuth account, or local server endpoint.
  </Step>

  <Step title="Validate the connection">
    Click **Validate** when available. For local providers, make sure the local server is running before validating.
  </Step>

  <Step title="Select Model">
    Choose a model from the dropdown. If the list is empty, fix credentials first, then refresh or reopen the provider settings.
  </Step>

  <Step title="Test with a short prompt">
    Send a short message before using the provider in long workflows, agents, or automations.
  </Step>
</Steps>

## Supported Providers

### Subscription Plans

Use these when you want EnConvo to connect through an existing consumer subscription instead of a token-billed developer API.

| Provider                          | Description                          | Link                                      |
| --------------------------------- | ------------------------------------ | ----------------------------------------- |
| **OpenAI Plus/Pro Subscription**  | Use your ChatGPT Plus or Pro account | [docs](/providers/openai_subscription)    |
| **Claude Pro/Max Subscription**   | Use your Claude Pro or Max account   | [docs](/providers/anthropic_subscription) |
| **Gemini Pro/Ultra Subscription** | Use your Google Gemini subscription  | [docs](/providers/google_subscription)    |

### Cloud API Providers

Use these when you want direct API billing, provider dashboards, and explicit API keys.

| Provider          | Description                             | Link                          |
| ----------------- | --------------------------------------- | ----------------------------- |
| **EnConvo Cloud** | Built-in LLM service with points system | [docs](/providers/enconvo)    |
| **OpenAI**        | GPT-4o, GPT-4, o1, o3 models            | [docs](/providers/openai)     |
| **Anthropic**     | Claude 4, Claude 3.5 Sonnet             | [docs](/providers/anthropic)  |
| **Google**        | Gemini 2.5, Gemini 3 Pro                | [docs](/providers/google)     |
| **DeepSeek**      | DeepSeek Chat, DeepSeek Reasoner        | [docs](/providers/deepseek)   |
| **Groq**          | Ultra-fast inference                    | [docs](/providers/groq)       |
| **xAI**           | Grok models                             | [docs](/providers/xai)        |
| **Mistral**       | Mistral Large, Mistral Small            | [docs](/providers/mistral)    |
| **Qwen**          | Alibaba's Qwen models                   | [docs](/providers/qwen)       |
| **Cohere**        | Command R models                        | [docs](/providers/cohere)     |
| **Perplexity**    | Sonar search models                     | [docs](/providers/perplexity) |

### Aggregators & Gateways

| Provider         | Description                                                 | Link                           |
| ---------------- | ----------------------------------------------------------- | ------------------------------ |
| **Hermes Agent** | Local or self-hosted agent runtime through Chat Completions | [docs](/providers/hermes)      |
| **OpenClaw**     | Self-hosted agent Gateway through Chat Completions          | [docs](/providers/openclaw)    |
| **OpenRouter**   | Access 100+ models via one API                              | [docs](/providers/openrouter)  |
| **Together AI**  | Open-source model hosting                                   | [docs](/providers/together-ai) |
| **Fireworks**    | Fast inference platform                                     | [docs](/providers/fireworks)   |

### Enterprise

| Provider         | Description                                                | Link                            |
| ---------------- | ---------------------------------------------------------- | ------------------------------- |
| **Azure OpenAI** | Microsoft's OpenAI service with deployment-based endpoints | [docs](/providers/azure-openai) |

### Speech Providers

| Provider   | Description                                    | Link                      |
| ---------- | ---------------------------------------------- | ------------------------- |
| **Soniox** | Speech-to-text for dictation and transcription | [docs](/providers/soniox) |

### Local Models

| Provider      | Description                                                | Link                         |
| ------------- | ---------------------------------------------------------- | ---------------------------- |
| **Ollama**    | Run local LLMs easily                                      | [docs](/providers/ollama)    |
| **LM Studio** | GUI for local GGUF models with an OpenAI-compatible server | [docs](/providers/lm-studio) |

## Common Settings

All providers share these configuration options:

| Setting                  | Description                                                                             |
| ------------------------ | --------------------------------------------------------------------------------------- |
| **Credentials**          | API key, OAuth account, subscription login, local token, or provider account connection |
| **Base URL / Endpoint**  | Required for local servers, gateways, Azure OpenAI, and OpenAI-compatible providers     |
| **Model Name**           | The specific model to use                                                               |
| **Temperature**          | Creativity level (0-2)                                                                  |
| **Reasoning / Thinking** | Optional reasoning controls for models that support it                                  |

<Note>
  Provider model lists and capabilities can change. EnConvo fetches model data dynamically when a provider supports it, and falls back to a known default when the provider cannot be reached.
</Note>

### Temperature Settings

| Level   | Value | Use Case                         |
| ------- | ----- | -------------------------------- |
| None    | 0     | Deterministic, factual responses |
| Low     | 0.5   | Balanced, slightly creative      |
| Medium  | 1     | Default, good balance            |
| High    | 1.5   | More creative                    |
| Maximum | 2     | Most creative, experimental      |

## Advanced Features

### Extended Thinking (Claude)

For Claude models with reasoning capabilities:

* **Disabled**: Standard response
* **Minimal**: 1,024 thinking tokens
* **Low**: 2,048 thinking tokens
* **Medium**: 5,120 thinking tokens
* **High**: 10,240 thinking tokens

### Reasoning Effort (OpenAI o-series)

For OpenAI reasoning models (o1, o3, o4):

* **Low**: Fast, economical
* **Medium**: Balanced
* **High**: More thorough reasoning

### Gemini Thinking

For Google Gemini 2.5+ models:

* **Disabled**: No thinking
* **Auto**: Model decides
* **Custom budgets**: 512 to 30,000 tokens

### Google Search Tool

Available for Gemini 2.5 models - grounds responses with real-time web content.

### URL Context Tool

Available for Gemini 2.5 models - allows including URLs for additional context.

## Validation Checklist

Use this checklist when any provider fails to respond:

<AccordionGroup>
  <Accordion title="Credentials">
    Confirm the API key, OAuth login, subscription connection, or local bearer token is current. Revoke and replace exposed API keys.
  </Accordion>

  <Accordion title="Billing and quota">
    Cloud providers may reject requests when billing is not configured, credits are exhausted, or rate limits are reached.
  </Accordion>

  <Accordion title="Base URL">
    Use the provider base URL, not a full endpoint path. For OpenAI-compatible services, the base URL usually ends in `/v1`.
  </Accordion>

  <Accordion title="Model access">
    Some models require account approval, region availability, paid billing, or a specific deployment name.
  </Accordion>

  <Accordion title="Local services">
    For Ollama, LM Studio, Hermes, and OpenClaw, make sure the local server is running and reachable from the same machine or LAN address used in EnConvo.
  </Accordion>
</AccordionGroup>

## Choosing a Provider

<AccordionGroup>
  <Accordion title="Best for general use">
    * **OpenAI GPT-4o**: Excellent all-around performance
    * **Anthropic Claude 3.5 Sonnet**: Great for writing and analysis
    * **Google Gemini 2.5 Pro**: Long context, multimodal
  </Accordion>

  <Accordion title="Best for coding">
    * **Anthropic Claude 3.5 Sonnet**: Excellent code understanding
    * **Mistral Codestral**: Code-specialized model
    * **DeepSeek Reasoner**: Strong for code analysis and debugging
    * **OpenAI GPT-4o**: Strong coding abilities
  </Accordion>

  <Accordion title="Best for speed">
    * **Groq**: Ultra-fast inference
    * **Fireworks AI**: Low-latency hosted open models
    * **Gemini Flash**: Fast cloud model for lightweight tasks
  </Accordion>

  <Accordion title="Best for privacy">
    * **Ollama**: Completely local
    * **LM Studio**: Local with GUI
  </Accordion>

  <Accordion title="Best for budget">
    * **EnConvo Cloud**: Pay with points
    * **DeepSeek**: Very affordable
    * **Groq Free Tier**: Free usage available
  </Accordion>
</AccordionGroup>