Introduction

oHallo is an autonomous B2B customer care platform. It receives customer messages across email, chat, WhatsApp, and voice channels, then resolves them by reasoning over your knowledge base, policies, and business systems — without requiring human intervention for routine queries.

The platform itself contains no business logic. Pricing, ordering, returns, quoting, and every other domain-specific operation live in MCP servers that you control. oHallo’s AI agents call tools on those servers to look up data, execute actions, and compose informed replies.

Two integration paths

REST API

Use the oHallo API to read and write platform data programmatically. Common use cases:

Sync conversations and messages into your own systems
Manage knowledge base entries and policies via CI/CD pipelines
Search and update contacts and accounts
Build custom dashboards or reporting tools
Trigger outbound conversations from external events

MCP Servers

Build MCP (Model Context Protocol) servers that expose your business systems as tools the AI can call. When a customer asks about an order, the agent calls get_order on your MCP server. When they need a quote, it calls create_quote.

You define what tools are available and what each tool does. The AI handles the conversation flow.

How it works

Customer message
    |
    v
Channel (email / chat / WhatsApp / voice)
    |
    v
AI Pipeline
    |-- Orchestrator: classifies intent, selects specialist agents
    |-- Specialist Agents: call MCP tools to gather data
    |-- Message Agent: composes the reply
    |
    v
MCP Server (you build this)
    |
    v
Your System (ERP, CRM, order management, etc.)

The AI pipeline runs automatically for every inbound message. Human agents can take over any conversation at any time, and the AI resumes when they release it.

What you can build

Custom integrations — Push conversation events into Slack, create tickets in Jira, or sync resolved conversations to your data warehouse.

CRM sync — Keep contacts and accounts in sync between oHallo and your CRM. oHallo auto-creates contact records from inbound messages; use the API to enrich them or link them to existing accounts.

Automated KB management — Bulk-import knowledge base entries from your documentation system. Approve or reject AI-proposed entries via the API. Keep your KB in sync with product changes.

Embedded widgets — Use the public KB widget API to embed a searchable help centre on your website, styled to match your brand.

Custom agent specialisation — Define specialist agents with custom instructions and bind them to specific MCP servers. A “returns agent” that knows your returns policy and can call create_return on your returns MCP server, for example.

Get started

Core Concepts Understand tenants, workspaces, channels, conversations, and MCP servers.

Authentication API keys, scopes, and security best practices.

API Conventions Base URL, pagination, errors, and rate limiting.

Build an MCP Server Connect oHallo to your business system.