Funding
Funding not disclosed
Founders
Product
Problem
Developers and organizations building AI-driven applications often struggle with non‑deterministic execution, lack of reliable event logging, and difficulty reproducing environments for testing, CI pipelines, and audits. Integrating external services (e.g., email, calendar) adds further complexity, especially when offline or repeatable builds are required.
Solution
Aligntrue offers a modular, open‑source platform that provides a deterministic kernel and host runtime for supervised AI execution. The kernel records events in an immutable log, uses deterministic envelopes, and supports idempotent outbox handling to ensure reliable side‑effects. A set of connectors (such as Gmail and Calendar) and reusable packs with UI blocks enable developers to build offline‑first applications that can be rebuilt from fixtures. The architecture is designed for repeatable CI pipelines and audit‑ready deployments, allowing teams to verify AI behavior across environments without reliance on external services.
Target Audience
Primary users are AI developers, data engineering teams, and enterprises that require reproducible, auditable AI workflows and integration with external services.
Features
- Receipts‑first kernel with core event log, deterministic envelopes, and rebuildable projections
- Idempotent outbox mechanism guaranteeing exactly‑once side‑effect execution
- Host runtime with governed safety classes for controlled AI dispatch
- Connectors for common services (e.g., Gmail, Calendar) to integrate external data sources
- Packs and UI blocks that facilitate rapid construction of offline‑first applications
- Fully open‑source MIT‑licensed codebase supporting offline operation with fixtures
- Designed for repeatable CI pipelines and audit‑trail generation