Jellyfish is the core engine of Truthful Intelligence, a lightweight AI library written in pure C with no external dependencies. Designed for embedded and trust-critical systems, it records every input, output, and decision as a cryptographically hashed block, forming a fully auditable chain of reasoning. With support for adaptive learning, memory-based inference, and automatic pruning, Jellyfish ensures transparent, verifiable AI behavior where every conclusion is traceable to its source.
-
Cross-Platform
Runs reliably on Windows, macOS, Linux, and embedded systems. -
Zero Dependencies
Written entirely in pure C for maximum portability, auditability, and minimal footprint. -
Blockchain-Inspired Memory
Every input, output, and decision is stored as a cryptographically hashed block, creating a tamper-resistant, traceable chain of logic. -
Self-Pruning Engine
Automatically removes obsolete or conflicting data, preserving clarity and consistency over time. -
Fully Auditable Reasoning
Supports forensic-level inspection of every decision, enabling ethical and transparent AI behavior. -
Modular and Configurable
Built to be embedded, extended, or customized for Truthful Intelligence applications across any platform.
To get started, ensure you have the following installed:
- Meson Build System: If you don’t have Meson
1.8.0
or newer installed, follow the installation instructions on the official Meson website. - Conan Package Manager: If you prefer using Conan, ensure it is installed by following the instructions on the official Conan website.
To add a git-wrap, place a .wrap
file in subprojects
with the Git repo URL and revision, then use dependency('fossil-jellyfish')
in meson.build
so Meson can fetch and build it automatically.
packages directly from a GitHub repository if it contains a valid conanfile.py
.
conan install git+https://github.com/fossillogic/fossil-jellyfish.git#v0.1.4 --name fossil_jellyfish --build=missing
Add the fossil-jellyfish.wrap
file in your subprojects
directory and include the following content:
[wrap-git]
url = https://github.com/fossillogic/fossil-jellyfish.git
revision = v0.1.4
[provide]
dependency_names = fossil-jellyfish
Note: For the best experience, always use the latest releases. Visit the releases page for the latest versions.
Jellyfish offers configurable options to tailor the build process to your needs:
- Running Tests: To enable testing, configure the build with
-Dwith_test=enabled
.
Example:
meson setup builddir -Dwith_test=enabled
The project is designed so that test cases serve two purposes:
- ✅ Unit Tests – validate the framework’s correctness.
- 📖 Usage Samples – demonstrate how to use these libraries through test cases.
This approach keeps the codebase compact and avoids redundant “hello world” style examples.
Instead, the same code that proves correctness also teaches usage.
This mirrors the Meson build system itself, which tests its own functionality by using Meson to test Meson.
In the same way, Fossil Logic validates itself by demonstrating real-world usage in its own tests via Fossil Test.
meson test -C builddir -v
Running the test suite gives you both verification and practical examples you can learn from.
For those interested in contributing, reporting issues, or seeking support, please open an issue on the project repository or visit the Fossil Logic Docs for more information. Your feedback and contributions are always welcome.