Agent integrations

​
Agent setup

Connect an MCP-compatible agent to your Orchestor workspace using OAuth. The hosted server URL is https://mcp.orchestor.io/mcp.

​
Choose your agent

Select an agent to get step-by-step setup instructions.

Filter by workflow:
Anthropic

Claude Code

Register MCP from your terminal and connect Claude Code to Orchestor.

View guide
OpenAI

Codex

Add the MCP server and connect your workspace with OAuth.

View guide
Cursor

Cursor

Add MCP to your editor to access Orchestor data while you work.

View guide
OpenCode

OpenCode

Configure the remote MCP server and authenticate to access your data.

View guide

​
Compare agents

Claude Code

Pricing
Subscription
Model
Single provider
Context
Project memory
Open source
No

Codex

Pricing
Hybrid
Model
Single provider
Context
Project memory
Open source
Yes

Cursor

Pricing
Subscription
Model
Multi-provider
Context
Indexed codebase
Open source
No

OpenCode

Pricing
BYOK
Model
Multi-provider
Context
Project memory
Open source
Yes

Skills and MCP support depends on the agent and configuration.

​
Understanding agents

Common types, concepts, and tradeoffs.

Workflow

Where the agent runs changes how you interact with it.

Terminal

Runs in a shell. Useful for automation, scripting, and CI pipelines.

IDE

An editor with integrated AI. Review visual diffs and edits across files.

Cloud

Runs on hosted infrastructure. Delegate asynchronous or long-running tasks.

Extension

Adds an agent to an existing editor while keeping your familiar setup.

Key concepts

The vocabulary you will encounter when comparing agents.

Skills

Reusable instructions and resources that teach an agent how to perform a task in a specific domain.

MCP

Model Context Protocol connects agents to external tools and data. Orchestor exposes workspace data through its MCP server.

Model flexibility

Some agents use only their vendor’s models. Others support multiple providers or your own API key (BYOK).

Context

The information available to an agent: the current conversation, persistent project instructions, or a searchable codebase.

Common tradeoffs

Decisions to consider when choosing an agent.

Cloudvs.Local

Cloud agents execute remotely; local agents execute on your machine. Local execution may still send context to a remote model, depending on the configuration.

Proprietaryvs.Open source

Proprietary agents follow the vendor’s license. Open-source agents make their source available to inspect and modify under their license.

Locked modelvs.BYOK

A fixed-model agent uses a specific provider. BYOK lets you configure your own API key for supported providers and models.

Sessionvs.Indexed codebase

Session context centers on the current conversation. An index makes repository files searchable beyond what is already in the conversation.