Documentation

Online

• API documentation: reference material for developing TORCS, plus the FAQ, Robot Tutorial and content-authoring guides.

Download for offline reading

Download documentation from SourceForge. Choose a release to get its prebuilt HTML documentation; no local generation is needed.

For robot (AI) developers

• TORCS Robot Tutorial: rewritten for 1.3.10 with a focus on learning and teaching. Build your own C++ racing driver with Linux and Windows setup instructions and complete checkpoint examples. Lessons cover steering, speed control, opponents and pit stops, with optional appendices on car setups, textures and multiple robots.
• The Simulated Car Racing Championship (SCR) extends TORCS with a server patch for controlling drivers through a network connection. Its competition software manual explains how to get started; the SCR Championship archive provides the software and examples. SCR is separate from the built-in C++ drivers covered by the Robot Tutorial.
• Published research on building TORCS drivers is collected on the Research page - a practical starting point for reinforcement learning approaches.

For track and car authors

• The source tree ships content tools: trackgen, accc, nfs2ac and texmapper.
• Track structure, XML descriptors and raceline generation are covered in the track manual. The expanded track toolchain documentation covers the supporting authoring tools, including map generation, elevation and banking workflows.
• The TORCS Content Creation and Toolchain video series demonstrates content workflows and the supporting tools.

Generate locally

• Linux: with Doxygen installed, run make doc from a configured TORCS source tree.
• Windows: install Doxygen 1.14.0 or newer and Graphviz, and add their bin directories to PATH. Open PowerShell or Command Prompt in the source directory containing TORCS.sln and run doxygen src/doc/torcsdoc.conf.

See the source package's README for full setup and output details.