Introduction

Origin is a Confidential Development Environment (CDE). It is a unified, browser-based platform where AI agents build and modify software under human direction, inside a cryptographically secure execution environment.

Engineers and teams working on high-value or sensitive codebases use Origin to accelerate development workflows. Internal benchmarking indicates faster development cycles for engineering teams using Origin, while security and compliance teams gain clearer visibility into how sensitive code is handled. Every workflow runs inside an environment designed for confidentiality, traceability, and controlled autonomy.

Note: Origin is purpose-built for teams that need both development velocity and verifiable security, from individual engineers working on proprietary code to regulated enterprises in finance, healthcare, government, and other security-critical sectors.

All Origin workspaces run inside a confidential environment. Powered by its proprietary OLLM AI Gateway, Origin gives you control over the assurance level applied to model inference specifically:

Standard model access for routine development work.
TEE-enabled models for higher-assurance inference, where enclave-backed confidentiality is required.

For confidential inference, Origin generates cryptographic attestation evidence tied to verified enclave execution, making confidentiality verifiable rather than a vendor claim. Inference activity is reviewable within the platform to support internal security workflows.

Built for regulated and security-critical environments, Origin combines hardware-backed encryption, verifiable isolation, and a privacy-first retention posture for customer content, making it suitable for financial institutions, healthcare organizations, public sector teams, and any engineering group handling high-value intellectual property.

This documentation covers how to integrate repositories, collaborate with agents, configure secure execution, and build software with verifiable autonomy across the full development lifecycle.

Core Capabilities

The following capabilities define how Origin operates as a Confidential Development Environment.

Confidential Computing & Attestation

TEE-enabled models run on hardware-backed confidential infrastructure. For qualifying inference requests, Origin generates cryptographic attestation records tied to verified enclave execution.

Unified Context Engine

Persistent context connects repositories, tasks, and research, enabling agents to maintain architectural awareness across files and sessions.

Agentic Development

Autonomous agents write, modify, and debug code with full traceability and reversible actions, under human direction throughout.

Online CDE & Live Preview

A secure, isolated browser-based environment with live preview, synchronized file operations, and runtime sandboxing. No local setup or editor plugins required.

Model Flexibility via OLLM

Select standard models for routine development or TEE-enabled models for higher-assurance inference, all through a single interface for model access, observability, and attestation-linked review.

Transparent Reasoning & Audit Logs

Inspect streaming agent reasoning, tool calls, and line-by-line file diffs. Usage and inference activity are reviewable within the platform.

Controlled Collaboration

Role-based permissions and team-based access controls ensure secure coordination across engineering, security, and compliance stakeholders.

Parallelized Workstreams

Run multiple agent tasks simultaneously while maintaining architectural coherence and centralized oversight.

Context Management

Attach or detach files, manage persistent project memory, and shape how agents understand your codebase across sessions.

Foundational Principles

Origin is built on three core principles. Provable Security & IP Protection is foundational.

1. Provable Security & IP Protection

Security is not a feature; it is the operating layer.

Hardware-backed confidential computing for sensitive projects and operations.
Cryptographic verification and attestation evidence for enclave execution.
Privacy-first retention posture: customer content is not used to train shared foundation models.
Well-suited for teams in regulated sectors as well as any engineering group handling sensitive or high-value code

Confidentiality in Origin is cryptographically verifiable, not policy-based.

2. Speed & Capacity

Autonomous agents accelerate delivery without increasing organizational overhead.

Internal benchmarking indicates faster development cycles for engineering teams using Origin.
Operate with expanded capacity without expanding headcount.
Coordinate multi-file and cross-service changes coherently.
Maintain continuity across large codebases and long-running initiatives.

Speed does not compromise control. Agents operate transparently under human direction.

3. True Full-Stack & Unified Context

Origin is designed for full-stack engineering, not isolated code generation.

Frontend, backend, infrastructure, and glue code in one environment.
Persistent context across repositories, sessions, and workflows.
Architectural awareness spanning services and dependencies.
Continuous reasoning that links planning and coding.

Rather than fragmenting research tools, IDEs, assistants, and compliance layers, Origin unifies them into a single, verifiable system.

Who Origin Is Built For

Origin is designed for engineering teams that require both velocity and verifiable security.

It is particularly suited for:

Engineers & independent developers: Ship sensitive or high-value software with the same confidentiality guarantees demanded by regulated enterprises.
AI-native startups: Protect proprietary models, datasets, and core IP during active development.
Regulated enterprises: Finance, healthcare, and government teams meeting audit obligations and third-party risk requirements.
Security-conscious SaaS teams: Maintain clear control boundaries over core intellectual property without slowing down delivery.

For teams operating under compliance requirements or IP risk constraints, Origin provides a development environment where autonomy, speed, and provable confidentiality coexist.