TesslateAI/OpenSail

What it solves

OpenSail is an open-source platform designed to turn recurring manual processes into durable, agentic software. It bridges the gap between simple AI experiments and production-ready automations by providing a control plane for managing a fleet of agents, connecting them to business systems, and ensuring human oversight through approval gates and budget limits.

How it works

OpenSail operates using a system of snapshot-backed workspaces (built on BtrFS) that provide persistent, portable environments for agents and apps. The platform uses a "Trigger -> Event -> Run -> Action -> Delivery" model to execute workflows. Users can define workflows in plain English, which are then converted into runnable systems with specific triggers (cron, webhooks, Slack), actions (agent tasks or app functions), and delivery targets.

Who it’s for

It is designed for developers, operators, founders, and professional teams (such as legal or support teams) who need to automate repetitive tasks—like lead follow-ups, billing reviews, or product feedback routing—while maintaining strict control over data, costs, and permissions.

Highlights

• Agentic Workflows: Build and share reusable agents that can research, update CRMs, and interact with business tools via MCP servers and connectors.
• Sandboxed Workspaces: Agents run in isolated environments with their own files, terminals, and containers, allowing for complex coding and system tasks.
• App Marketplace: Turn workspaces into versioned, installable apps with typed actions and views that can be shared across a team.
• Human-in-the-Loop: Integrated approval boundaries that pause runs for human review before risky actions are executed.
• Autonomous Builders: Built-in agents (@agent-builder and @automation-builder) that can draft new agents and workflows from chat descriptions.
• Full IDE Integration: A built-in Monaco editor and terminal where humans and agents collaborate on the same codebase with reviewable diffs.