Application-owned
Identity, policy, domain data and execution stay with the host application.
Independent AI engineering
About Loom AI Labs
Loom AI Labs is the home of open-source AI foundations, managed deployment infrastructure, runnable engineering proof and implementation-linked research into dependable AI-enabled software.
Maintained by Mahmoud Ashraf Algammal with open-source contributors
Mission
Make AI behavior legible inside real applications.
The work focuses on the space between a model API and a dependable product: trusted application context, current data, retrieval evidence, governed actions, visible limitations and reusable user experience.
Loom AI Labs does not treat prompts as a substitute for identity, authorization, source-of-truth data or business execution.
Engineering principles
A small set of rules shapes every layer.
Explicit contracts
Context, evidence, clarification and actions cross typed, inspectable boundaries.
Proof before claims
Live demos and public implementation support claims; limitations remain visible.
Composable adoption
Teams can adopt one capability without relocating their whole application.
Source in view
Products and runnable proof link back to the implementation they describe.
Bounded language
Prototype, evidence and maturity terms match what public artifacts establish.
One connected body of work
Products, experiments and research have different jobs.
- 01 Engineering question
- 02 Reusable foundation
- 03 Managed deployment
- 04 Runnable proof
Products
Open-source foundations and a managed platform with explicit ownership boundaries.
Experiments
Bounded applications that make one technical capability inspectable.
Research
Engineering investigations linking questions, implementations and limitations.
Maintainer identity
Mahmoud Ashraf Algammal
Loom AI Labs is independently maintained. Contributions are reviewed against the same application-boundary, evidence and source-linking principles used across the public work.
Open-source model
Public foundations, inspectable change.
AI Fabric Framework and AI Fabric Chat UI are released under Apache 2.0. Issues and pull requests belong in the relevant public repository, close to code and verification.
Contribution path
Start with a reproducible boundary.
Bring an issue, an integration gap or a bounded experiment with enough context for another engineer to inspect it.