AI Coding allows us to build complete features extremely fast. Many of these features directly solve specific customer problems. However, merging them into the core FeatBit product requires strict review and testing to maintain enterprise-level reliability. This can slow down the existing main development process, and delay delivering valuable features to customers.
This situation is not compatible with FeatBit’s commitment to Customer Success, particularly in today’s fast-moving AI innovation landscape. To balance speed and stability, we created the FeatBit Experimental Program repository.
Experimentals is a collection of independent feature applications generated or accelerated by AI, designed to solve real customer pain points immediately.
These are fully usable standalone projects, not concept demos. They include:
- Implementation guidance leveraging FeatBit’s existing features, APIs, and data structures
- Fully runnable reference implementations with source code and documentation
Customers can deploy and use these applications right away without writing code themselves. If the applications not fully meet their needs, customers can modify or extend them as desired based on the provided tutorials and source code.
Customers do not need to wait for a major product release—Experimentals provides an immediate, working solution.
If customers need more specific functionality, they can modify the open codebase based on our examples and tutorials.
Real usage helps us validate demand, refine design, and collect feedback before merging a stable version into the core FeatBit platform.
Experimental innovations are isolated, so we can keep FeatBit’s production-level reliability uncompromised.
Build fast. Deliver value immediately.
Validate with real customers.
Merge only when proven.
Experimentals lets us maximize innovation speed while ensuring the FeatBit core remains secure, reliable, and trusted—fully aligned with our ultimate mission: Customer Success.
Each folder in this repository represents an independent project that addresses a specific customer need.
Each project folder may include:
- A markdown file include the definition of the problem, solution approach, and usage instructions.
- An application with code source to implement the solution.
The most of code are generated by AI Coding tools, but we have tested and refined them to ensure they work as expected.
- Truncate Feature Flag Insights Data: Demonstrates how to truncate feature flag insights data from the database to manage storage and performance. [PostgreSQL Example Included, MongoDB and Clickhouse Coming Soon]