ChanlChanl
EngineeringPara Engineering Leads & Platform Teams

You wouldn't ship code without tests. Why ship AI agents without them?

82% of AI agent teams have no testing pipeline. No staging environment for conversations. No way to catch regressions before users do. Chanl gives you the testing and observability layer your AI agents need.

Comienza GratisVer Demo

API-first, not dashboard-first

RESTful APIs, TypeScript SDK, and MCP protocol. Build agents, manage tools, run tests, and deploy — all programmatically. No clicking through UIs.

Connect any tool or API

HTTP endpoints, JavaScript functions, and code actions your agents can call during live conversations. Connect your CRM, database, or internal service in minutes.

Debug with traces, not guesswork

Every conversation transcribed, every tool call traced with inputs/outputs/latency, every response scored. Find the root cause in minutes, not days.

Developer Experience

SDKs and APIs that respect your time

TypeScript SDK with full type safety. REST APIs with Swagger docs. MCP protocol for Claude Desktop and AI tools. Integrate into your existing stack — not the other way around.

  • TypeScript/JavaScript SDK with full type safety
  • REST API with Swagger documentation
  • MCP protocol for Claude Desktop and AI tools
  • Webhooks for real-time event notifications
system-prompt.md
v2.3 → v2.4
# System Prompt
You are a helpful customer support agent.
- Be polite and professional
+ Be warm, empathetic and professional
- Response time target: 5 seconds
+ Response time target: 3 seconds
Always verify customer identity.
+ Ask for preferred contact method.
Tests
All passing
CSAT Impact
+2.4%
Deployed
2 min ago
Extensibility

Your agents, your tools, your data

Build custom tools that let AI agents interact with your systems. RAG-powered knowledge bases, persistent memory, and prompt templates with variable substitution.

  • Custom tool builder with HTTP, JS, and code actions
  • Knowledge base with document upload and RAG search
  • Agent memory that persists across conversations
  • Prompt templates with variable substitution
Connected Integrations12 active
SalesforceSalesforce
SlackSlack
GoogleGoogle
StripeStripe
HubSpotHubSpot
IntercomIntercom
ZapierZapier
ShopifyShopify
GitHubGitHub
JiraJira
GmailGmail
PostgreSQLPostgreSQL
Observability

When it breaks at 2am, you'll know why

Full conversation transcripts, tool call traces, and automated scoring. Search across thousands of conversations. Find the regression before your users tweet about it.

  • Searchable transcripts with speaker turns and timestamps
  • Tool call tracing with inputs, outputs, and latency
  • Automated quality scoring with custom scorecards
  • Real-time analytics and alerting
chanl-cli
$ chanl scenarios run-all --agent agent_prod --min-score 80
Batch Scenario Results
────────────────────────────────────────────────────
ScenarioStatusScoreTimeResult
Billing disputecompleted96%8.2sPASS
Tech support triagecompleted74%12.1sFAIL
Account upgradecompleted92%6.7sPASS
VIP escalationcompleted88%14.3sPASS
Confused usercompleted67%9.8sFAIL
Summary
Total: 5
Passed: 3
Failed: 2
Average Score: 83%
2 of 5 scenarios failed
60%

Preguntas Frecuentes

Ship AI agents like you ship software

API-first platform with testing, tools, and observability. Free tier — no sales call needed.

Comienza GratisHablar con Ventas