Custom LLMs

Custom LLMs: bring your own AI models and API keys.

Bring OpenAI, Anthropic, Google, Azure, AWS, OpenRouter, xAI, DeepSeek, Groq, Cohere, Ollama, or your own model APIs. Route requests by cost, latency, or capability — with fallbacks, monitoring, and governance built in.

Get Started For Free Chat on WhatsApp

Or start with $150 of credit, enough for a first month of Growth

  • OpenAI
  • Anthropic
  • Google Gemini
  • Azure OpenAI
  • AWS Bedrock
  • OpenRouter
  • xAI
  • DeepSeek
  • Groq
  • Cohere
  • Ollama

Flexible Model Integration

Choose the approach that fits your stack — providers, routing, and governance.

  • Connect External APIs

    Bring your existing model providers into MonoChat via credentials and validate connections.

    • API key management
    • Secure credential storage
    • Connection validation
    • Usage monitoring

  • Model Registry

    Manage multiple models per workspace and assign them to channels, agents, and scopes.

    • Provider + model catalog
    • Per-workspace isolation
    • Environment separation
    • Versioned configs

  • Routing & Fallbacks

    Route by cost, latency, capability, or language — with safe fallbacks and A/B tests.

    • Intelligent routing rules
    • Fallback chains
    • A/B testing
    • Performance comparison

Supported API Providers

Connect your preferred providers without locking into a single vendor.

  • OpenAI

    General-purpose language, vision, and audio model families for assistants, search, and content generation.

    • Chat & assistant models
    • Multimodal reasoning models
    • Embeddings & tools

  • Anthropic

    Safety-focused assistant and reasoning model families for enterprise-grade conversations.

    • Assistant-style models
    • Long-context reasoning models
    • Enterprise safety tooling

  • Google

    Google AI model families for search, conversation, and multimodal experiences.

    • Search & conversation models
    • Multimodal models
    • Enterprise ML services

  • Azure OpenAI

    Enterprise OpenAI services with Azure compliance, observability, and governance.

    • Managed foundation models
    • Enterprise security & compliance
    • Monitoring & observability

  • AWS Bedrock

    Fully managed foundation models on AWS, running inside your own cloud account and region.

    • Multiple model families
    • Runs in your AWS account
    • IAM & VPC integration

  • OpenRouter

    One API key for hundreds of models from many providers, with unified billing and automatic provider fallbacks.

    • Hundreds of models via one API
    • Open-source & proprietary models
    • Provider fallbacks & price routing

  • xAI

    Grok model family for fast, real-time-aware conversation and reasoning.

    • Grok chat models
    • Reasoning models
    • Vision support

  • DeepSeek

    Cost-efficient chat and reasoning models with strong coding and math performance.

    • Chat models
    • Reasoning models
    • Low cost per token

  • Groq

    Ultra-low-latency inference for popular open models, ideal for real-time chat.

    • Very fast inference
    • Open-source models
    • Real-time conversations

  • Cohere

    Enterprise language models built for retrieval, search, and multilingual business use.

    • Command chat models
    • Embeddings & rerank
    • Multilingual support

  • Ollama

    Run open-source models on your own servers and keep every conversation on your infrastructure.

    • Self-hosted models
    • Data stays on-premise
    • No per-token fees

  • Custom APIs

    Your own in-house or domain-specific model APIs, fully controlled by your team.

    • Internal model gateways
    • Domain-specific models
    • Private inference endpoints

Easy Integration Process

Get your custom models up and running with a clean, repeatable setup.

  1. Add API Credentials

    Securely add your provider keys and restrict access by role.

  2. Configure Routing

    Set model preferences, fallbacks, and rules per channel, agent, or workflow.

  3. Test & Deploy

    Run tests, monitor performance, and roll out safely to production.

Enterprise-Grade Security

Protect credentials and enforce governance for every model and workspace.

  • Credential Encryption

    All provider credentials are encrypted at rest and in transit.

  • Access Control

    Granular permissions with role-based access control (RBAC).

  • Audit Logging

    Full audit trail of model configuration changes and usage.

  • Workspace Isolation

    Data and model configurations are isolated per workspace.

Industry-Specific Models

Examples of domain-focused model stacks teams connect to MonoChat.

  • E-commerce

    • Product recommendation engine
    • Customer sentiment analysis
    • Price optimization model
    • Inventory prediction

  • Healthcare

    • Medical intake assistant
    • Patient triage system
    • Drug interaction checker
    • Symptom analysis

  • Finance

    • Fraud detection system
    • Credit scoring model
    • Risk assessment tool
    • Compliance summaries

  • Education

    • Personalized learning path
    • Student performance predictor
    • Content recommendation
    • Automated grading

FAQ

Custom LLM questions

Quick answers to the most common questions.

What does "bring your own LLM" mean in MonoChat?

You connect your own API credentials for the model providers you already use. MonoChat calls the models through your account, so you keep your pricing and data agreements, and there is no AI markup.

Which LLM providers can I connect?

OpenAI, Anthropic, Google, Azure OpenAI, AWS Bedrock, OpenRouter, xAI, DeepSeek, Groq, Cohere and Ollama, plus your own custom model APIs.

Can different assistants or channels use different models?

Yes. The model registry lets you manage several models per workspace and assign them to channels, agents and scopes. Routing rules can choose a model by cost, latency, capability or language, with fallback chains if a provider fails.

Can I use self-hosted or private models?

Yes. You can run open-source models on your own servers with Ollama, or connect an in-house model through a custom API endpoint.

How are my provider API keys protected?

Provider credentials are encrypted at rest and in transit. Role-based access control limits who can use them, changes are recorded in an audit log, and model configurations are isolated per workspace.

Keep exploring

First month of Growth on us

Ready to Connect Your Models?

Add credentials, configure routing, and run multi-model assistants in MonoChat — with governance and monitoring.

Start with $150 credit Chat on WhatsApp

One code per business • 30 days to redeem • No card needed

$150 to get started.

Claim