Loading…

Developers

Every agent is an API. Build on ZeaIQ your way.

Standard protocols in, standard protocols out. Call governed agents from any app, expose them to any MCP client, and extend them with your own APIs and tools.

Book a demo Contact sales

OpenAI-compatible API

Use agents anywhere you use a model

Point any OpenAI client at ZeaIQ and address agents or teams as models. Streaming, model listing and legacy completions are supported.

POST /v1/chat/completions, GET /v1/models, /v1/completions
Models: agent:{uuid} or team:{uuid}
API key scope agents:openai_compat, as Bearer or X-Api-Key

Read the docs (sign in)

OpenAI-compatible

chat.py

from openai import OpenAI

# Any OpenAI client works — point it at ZeaIQ.
client = OpenAI(
    base_url="https://your-instance.zeaiq.ai/v1",
    api_key=ZEAIQ_API_KEY,  # scope: agents:openai_compat
)

reply = client.chat.completions.create(
    model="agent:6d85c1fb-0dc7-4529-a793-74d7e1d7c7d4",
    messages=[{"role": "user", "content": "Summarize today's payment exceptions"}],
    stream=True,
)

MCP server

Agents and workflows as MCP tools

Switch on Expose via MCP and any MCP client — IDEs, desktop assistants, other agents — can call your governed agents and workflows as tools.

Streamable HTTP, stateless, one server per organization
Agents appear as ask_* tools, workflows as run_* tools
Policy rules can target the mcp surface specifically

Read the docs (sign in)

MCP server

mcp.json

{
  "mcpServers": {
    "zeaiq": {
      "url": "https://your-instance.zeaiq.ai/api/mcp",
      "headers": {
        "Authorization": "Bearer <api-key with mcp:invoke>"
      }
    }
  }
}
// Exposed agents appear as ask_* tools,
// exposed workflows as run_* tools.

A2A protocol

Agent-to-agent, out of the box

Every agent and team gets an A2A agent card and a JSON-RPC endpoint, so other agent platforms can discover and call them.

Agent cards at /.well-known/agent-card.json
message/stream over JSON-RPC with A2A task states
Scopes agents:a2a and agent_teams:a2a

Read the docs (sign in)

A2A

a2a.http

GET  /a2a/agents/<id>/.well-known/agent-card.json

POST /a2a/agents/<id>/?key=<api-key>
Content-Type: application/json

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "message/stream",
  "params": {
    "message": {
      "role": "user",
      "parts": [{ "kind": "text", "text": "Summarize Q1 pipeline by region." }]
    }
  }
}

Embedding

A branded assistant on any website

Drop in one script tag for a floating widget or full-page chat. Control it from your own buttons with a small JavaScript API.

ZeaIQWidget.open(), ask(), close(), toggle()
Per-page overrides: color, size, alignment, launcher
iframe embed and public chatbot links with QR codes

Read the docs (sign in)

Website widget

index.html

<script
  src="https://your-instance.zeaiq.ai/widget.js"
  data-agent-id="<agentId>"
  data-key="<api-key>"
  data-api="https://your-instance.zeaiq.ai/api"
  data-launcher="bar"
></script>

<button onclick="ZeaIQWidget.ask('How do I open an account?')">
  Ask our assistant
</button>

Webhooks

Start workflows from any system

Each webhook trigger gets its own receiver URL and key, so external systems can start workflows with a single POST.

Per-trigger API keys with optional IP allowlists
Payload becomes the workflow input
Combine with schedules, folder watch and app events

Read the docs (sign in)

Webhook trigger

trigger.sh

# Start a workflow from any system
curl -X POST \
  https://your-instance.zeaiq.ai/api/webhooks/wf_<workflowId>/trigger_<triggerId> \
  -H "X-Api-Key: $TRIGGER_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "event": "invoice_received", "document_url": "https://…" }'

Extend

Bring your own tools, keep the guardrails.

Everything you add runs under the same approvals, policy rules, credential vault and audit as built-in capabilities.

OpenAPI tools

Import a spec and every operation becomes a tool your agents can call, with secrets from the org vault.

MCP client

Connect external MCP servers over HTTP or stdio and use their tools inside agents.

HTTP webhooks

Wrap any internal endpoint as a tool with typed inputs and optional human approval.

Skills

Ship procedures as Markdown with resource files and Python, Shell or JavaScript scripts.

Code interpreter

Sandboxed Python for analysis, with code, output, figures and files in the chat.

Workflows as code

Import and export workflows as YAML in the Microsoft Agent Framework declarative style.

API keys

Least privilege, per integration

Issue a key per client or team with only the scopes it needs, restrict it by origin or host, and rotate or revoke it without affecting anything else.

11 scopes, from agents:openai_compat to mcp:invoke
Origin and host restrictions for browser keys
Rotate and revoke at any time
REST API with OpenAPI documentation
Every call lands in Agent Ops and the audit log

Documentation

50 guides in English, French, Spanish, German, Chinese and Arabic — with an Ask AI assistant. Sign in to your workspace to read them.

Open the docs (sign in)

Build something your users will love.

Get a sandbox organization and an API key, and talk to our engineers about your architecture.

Request a sandbox Book a demo

Loading…

Developers

Every agent is an API. Build on ZeaIQ your way.

Standard protocols in, standard protocols out. Call governed agents from any app, expose them to any MCP client, and extend them with your own APIs and tools.

Book a demo Contact sales

OpenAI-compatible API

Use agents anywhere you use a model

Point any OpenAI client at ZeaIQ and address agents or teams as models. Streaming, model listing and legacy completions are supported.

POST /v1/chat/completions, GET /v1/models, /v1/completions
Models: agent:{uuid} or team:{uuid}
API key scope agents:openai_compat, as Bearer or X-Api-Key

Read the docs (sign in)

OpenAI-compatible

chat.py

from openai import OpenAI

# Any OpenAI client works — point it at ZeaIQ.
client = OpenAI(
    base_url="https://your-instance.zeaiq.ai/v1",
    api_key=ZEAIQ_API_KEY,  # scope: agents:openai_compat
)

reply = client.chat.completions.create(
    model="agent:6d85c1fb-0dc7-4529-a793-74d7e1d7c7d4",
    messages=[{"role": "user", "content": "Summarize today's payment exceptions"}],
    stream=True,
)

MCP server

Agents and workflows as MCP tools

Switch on Expose via MCP and any MCP client — IDEs, desktop assistants, other agents — can call your governed agents and workflows as tools.

Streamable HTTP, stateless, one server per organization
Agents appear as ask_* tools, workflows as run_* tools
Policy rules can target the mcp surface specifically

Read the docs (sign in)

MCP server

mcp.json

{
  "mcpServers": {
    "zeaiq": {
      "url": "https://your-instance.zeaiq.ai/api/mcp",
      "headers": {
        "Authorization": "Bearer <api-key with mcp:invoke>"
      }
    }
  }
}
// Exposed agents appear as ask_* tools,
// exposed workflows as run_* tools.

A2A protocol

Agent-to-agent, out of the box

Every agent and team gets an A2A agent card and a JSON-RPC endpoint, so other agent platforms can discover and call them.

Agent cards at /.well-known/agent-card.json
message/stream over JSON-RPC with A2A task states
Scopes agents:a2a and agent_teams:a2a

Read the docs (sign in)

A2A

a2a.http

GET  /a2a/agents/<id>/.well-known/agent-card.json

POST /a2a/agents/<id>/?key=<api-key>
Content-Type: application/json

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "message/stream",
  "params": {
    "message": {
      "role": "user",
      "parts": [{ "kind": "text", "text": "Summarize Q1 pipeline by region." }]
    }
  }
}

Embedding

A branded assistant on any website

Drop in one script tag for a floating widget or full-page chat. Control it from your own buttons with a small JavaScript API.

ZeaIQWidget.open(), ask(), close(), toggle()
Per-page overrides: color, size, alignment, launcher
iframe embed and public chatbot links with QR codes

Read the docs (sign in)

Website widget

index.html

<script
  src="https://your-instance.zeaiq.ai/widget.js"
  data-agent-id="<agentId>"
  data-key="<api-key>"
  data-api="https://your-instance.zeaiq.ai/api"
  data-launcher="bar"
></script>

<button onclick="ZeaIQWidget.ask('How do I open an account?')">
  Ask our assistant
</button>

Webhooks

Start workflows from any system

Each webhook trigger gets its own receiver URL and key, so external systems can start workflows with a single POST.

Per-trigger API keys with optional IP allowlists
Payload becomes the workflow input
Combine with schedules, folder watch and app events

Read the docs (sign in)

Webhook trigger

trigger.sh

# Start a workflow from any system
curl -X POST \
  https://your-instance.zeaiq.ai/api/webhooks/wf_<workflowId>/trigger_<triggerId> \
  -H "X-Api-Key: $TRIGGER_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "event": "invoice_received", "document_url": "https://…" }'

Extend

Bring your own tools, keep the guardrails.

Everything you add runs under the same approvals, policy rules, credential vault and audit as built-in capabilities.

OpenAPI tools

Import a spec and every operation becomes a tool your agents can call, with secrets from the org vault.

MCP client

Connect external MCP servers over HTTP or stdio and use their tools inside agents.

HTTP webhooks

Wrap any internal endpoint as a tool with typed inputs and optional human approval.

Skills

Ship procedures as Markdown with resource files and Python, Shell or JavaScript scripts.

Code interpreter

Sandboxed Python for analysis, with code, output, figures and files in the chat.

Workflows as code

Import and export workflows as YAML in the Microsoft Agent Framework declarative style.

API keys

Least privilege, per integration

Issue a key per client or team with only the scopes it needs, restrict it by origin or host, and rotate or revoke it without affecting anything else.

11 scopes, from agents:openai_compat to mcp:invoke
Origin and host restrictions for browser keys
Rotate and revoke at any time
REST API with OpenAPI documentation
Every call lands in Agent Ops and the audit log

Documentation

50 guides in English, French, Spanish, German, Chinese and Arabic — with an Ask AI assistant. Sign in to your workspace to read them.

Open the docs (sign in)

Build something your users will love.

Get a sandbox organization and an API key, and talk to our engineers about your architecture.

Request a sandbox Book a demo

Every agent is an API.Every agent is an API. Build on ZeaIQ your way.Build on ZeaIQ your way.

Use agents anywhere you use a model

Agents and workflows as MCP tools

Agent-to-agent, out of the box

A branded assistant on any website

Start workflows from any system

Bring your own tools,Bring your own tools, keep the guardrails.keep the guardrails.

OpenAPI tools

MCP client

HTTP webhooks

Skills

Code interpreter

Workflows as code

Least privilege, per integration

Documentation

Build somethingBuild something your users will love.your users will love.

Every agent is an API.Every agent is an API. Build on ZeaIQ your way.Build on ZeaIQ your way.

Use agents anywhere you use a model

Agents and workflows as MCP tools

Agent-to-agent, out of the box

A branded assistant on any website

Start workflows from any system

Bring your own tools,Bring your own tools, keep the guardrails.keep the guardrails.

OpenAPI tools

MCP client

HTTP webhooks

Skills

Code interpreter

Workflows as code

Least privilege, per integration

Documentation

Build somethingBuild something your users will love.your users will love.

Every agent is an API. Build on ZeaIQ your way.

Bring your own tools, keep the guardrails.

Build something your users will love.

Every agent is an API. Build on ZeaIQ your way.

Bring your own tools, keep the guardrails.

Build something your users will love.