byLLM Tutorials and Examples#
Tutorials#
-
RPG Game Level Genaration
A Tutorial on building an AI-Integrated RPG Game using byLLM.
-
Fantacy Trading Game
A text-based trading game where non-player characters are handled using large language models. Tool calling is used for game mechanics such as bargaining at shops.
-
AI-Powered Multimodal MCP Chatbot
This Tutorial shows how to implement an agentic AI application using the byLLM package and object-spatial programming. MCP integration is also demonstrated here.
Examples#
This section collects the example byllm programs bundled in jac-byllm/examples/. Examples are grouped by type. For each example the source is shown in a tab so you can quickly inspect the code.
Core Examples#
Small, focused examples that show common byLLM patterns for integrating LLMs in Jac programs.
Repository location: jac-byllm/examples/core_examples
Vision / Multimodal examples#
Examples that demonstrate multimodal usage (images and video) with byLLM and vision-capable LLMs. Accompanying media files live alongside the Jac code in the repo.
Repository location: jac-byllm/examples/vision
Vision / Multimodal examples (code)
Vision-enabled examples that combine image/video inputs with byLLM workflows. Only the Jac code files are shown below; accompanying media are in the examples folder (e.g. person.png, receipt.jpg, mugen.mp4).
Tool-calling examples#
Examples showing how to orchestrate external tools (APIs, search, or internal tool servers) from Jac/byLLM and how to coordinate multi-agent workflows.
Repository location: jac-byllm/examples/tool_calling
Tool-calling examples (code)
Examples that demonstrate calling external tools, tool orchestration, or multi-agent interactions.
Agentic AI examples#
Small agentic patterns and lightweight multi-step reasoning examples (multi-turn planning, simple agents). These live under the agentic_ai examples folder.
Repository location: jac-byllm/examples/agentic_ai
Agentic AI examples (code)
Examples that demonstrate small agentic behaviors and light-weight multi-step reasoning.
Microbenchmarks#
Short, single-purpose microbenchmarks for probing model behavior and performance on targeted tasks. Used for evaluations done in the MTPm paper.
Repository location: jac-byllm/examples/microbenchmarks