# Chanl

> Chanl is the infrastructure layer for AI agents in production. Build agents with tools, memory, and knowledge — test them with AI-powered scenarios before deployment — monitor every conversation with automated scorecards. Works with any voice, chat, or messaging platform.

Chanl is not an orchestration platform (like VAPI, Retell, or Bland). It is the backend infrastructure that makes AI agents reliable in production: persistent memory across channels, real tool execution via MCP, automated testing with AI personas, and conversation-level observability.

## What Chanl Does

Chanl has three pillars:

### Build (Agent Infrastructure)
- Configure AI agents once, deploy across voice, chat, SMS, and messaging
- Connect tools via MCP protocol, OpenAPI specs, or custom JavaScript/Python
- RAG-powered knowledge base with automatic chunking, embedding, and retrieval
- Persistent memory that follows customers across every interaction and channel
- Prompt version control with A/B testing and safe rollback

### Connect (Multi-Channel Deployment)
- One agent works on VAPI, ElevenLabs, Pipecat, LiveKit, OpenAI Realtime
- MCP server gives any AI platform access to Chanl's tools, memory, and knowledge
- Import existing agents from VAPI, ElevenLabs, or Bland with one API call
- Deploy to voice, web chat, SMS, WhatsApp, Slack from the same agent config

### Monitor (Testing & Observability)
- AI-powered test scenarios with 25+ configurable personas
- Automated scorecards grade every conversation on accuracy, tone, compliance, resolution
- Real-time analytics dashboards with conversation intelligence
- Live call monitoring with quality gates and instant alerts
- Source attribution traces every agent answer back to the document it came from

## Key Differentiators

- **MCP-native tool execution**: Agents use real tools (APIs, databases, CRMs) during conversations via Model Context Protocol — not just prompt-based responses
- **Cross-channel memory**: Persistent memory works across voice, chat, and SMS — customers never repeat themselves
- **AI-powered testing**: Test agents with realistic AI personas before customers interact with them
- **Automated scoring**: Every conversation gets graded automatically — no manual QA sampling
- **Platform-agnostic**: Works with any orchestration platform (VAPI, Retell, Bland, Pipecat, LiveKit, ElevenLabs)

## Pricing

- **Free**: $0/month — 3 agents, 5 tools, 1 knowledge base, MCP access, 50 test scenarios/month
- **Lite**: $29/month — 10 agents, 25 tools, 3 knowledge bases, 500 test scenarios/month
- **Startup**: $149/month — 50 agents, unlimited tools, 10 knowledge bases, unlimited scenarios, priority support
- **Business**: $299/month — unlimited agents, SSO, custom integrations, dedicated support, SLA

All plans include MCP access, persistent memory, and automated scorecards.

## Getting Started

- Sign up free: https://app.channel.tel/signup
- Documentation: https://www.channel.tel/docs
- Feature overview: https://www.channel.tel/features
- Pricing details: https://www.channel.tel/pricing
- Contact sales: https://www.channel.tel/contact

## Features

- AI Agents: https://www.channel.tel/features/agents
- AI Testing & Scenarios: https://www.channel.tel/features/scenarios
- Automated Scorecards: https://www.channel.tel/features/scorecards
- Analytics & Insights: https://www.channel.tel/features/analytics
- Live Call Monitoring: https://www.channel.tel/features/monitoring
- Prompt Management: https://www.channel.tel/features/prompts
- Agent Tools (MCP + OpenAPI): https://www.channel.tel/features/tools
- Persistent Memory: https://www.channel.tel/features/memory
- Model Context Protocol: https://www.channel.tel/features/mcp
- Knowledge Base (RAG): https://www.channel.tel/features/knowledge

## Solutions

- Contact Centers: https://www.channel.tel/solutions/contact-centers
- Sales Teams: https://www.channel.tel/solutions/sales
- Customer Success: https://www.channel.tel/solutions/customer-success

## FAQ

### What is Chanl?
Chanl is the backend infrastructure for AI agents in production. It provides tools, memory, knowledge, testing, and observability for AI agents deployed on any platform.

### How is Chanl different from VAPI or Retell?
VAPI and Retell are orchestration platforms — they handle the conversation layer (STT, TTS, LLM). Chanl handles everything else: persistent memory, real tool execution via MCP, automated testing with AI personas, and conversation-level scoring. Chanl works WITH orchestration platforms, not instead of them.

### Does Chanl replace my current voice AI platform?
No. Chanl works alongside your existing platform. If you use VAPI for voice, Chanl adds tools, memory, knowledge, and testing. You can also use Chanl's built-in Pipecat-based voice without any external platform.

### What is MCP and why does it matter?
MCP (Model Context Protocol) is an open standard that lets AI agents use real tools. Chanl's MCP server gives any AI platform access to your agent's tools, memory, and knowledge base through a single protocol.

### Can I test my AI agents before deploying?
Yes. Chanl's scenario engine lets you run AI-powered test personas against your agents. Configure persona traits (angry customer, confused user, edge case scenarios), run automated test suites, and score results with customizable scorecards.

### What does automated scoring do?
Every conversation your agent has gets automatically graded on dimensions you define: accuracy, tone, empathy, compliance, resolution success, response time. No manual QA sampling needed.

### Is there a free tier?
Yes. The free plan includes 3 agents, 5 tools, 1 knowledge base, MCP access, and 50 test scenarios per month. No credit card required.

### What platforms does Chanl integrate with?
Chanl works with VAPI, ElevenLabs, Pipecat, LiveKit, OpenAI Realtime, Retell, Bland, Twilio, and any platform that supports MCP or webhook integrations.