Capstone: Your TARS, Your Way and the Six-Month Showcase
This is the finish line of the whole program. Your TARS is printed, wired, assembled, and alive. Now you make it yours, give it a personality and a behavior no one else's has, polish it until it feels finished, write up what you built, and share it. Six months ago you uploaded a blink to a microcontroller; today you have a walking, talking, AI-powered robot on your desk, and it is time to show what you made.
Make it yours
A great capstone is personal. Two people who follow the same TARS-AI build should end up with two different robots, because you shape the personality, the voice, and the behavior. Adjust the prompt that gives the language model its character, choose a voice, and design gestures and responses that reflect the TARS you wanted to build. This is where the project stops being an assembly guide and becomes your creation, the same way every capstone this program has ended, with you putting your own stamp on the work.
A signature behavior that shows off the whole journey
Beyond the standard conversation and movement, add at least one behavior that is distinctly yours, and the best ones weave several months together. It might recognize you by sight and adjust the lights, combining Month 3 vision with Month 4 home control. It might tie a gesture to a spoken line using a Month 1 state machine and this month's motion programming. It might do a respectful, consented patrol-and-report drawing on Month 5. A signature behavior that pulls threads from across the six months is what turns a demo into a real demonstration that you can integrate skills, not just follow steps.
Calibrate, polish, and document
A finished robot is more than a working one. Spend time calibrating servo home positions and motion so movements land correctly and look smooth, tuning speech timing so conversation feels natural, and refining the personality so replies fit the character. That small polish, a steadier gait, a better voice, a snappier response, is the difference between a demo that limps and one that delights. Then capture it: write up your build, the parts, the choices you made, and the problems you solved, and record a short demonstration of TARS moving, talking, and doing its signature behavior. That becomes a portfolio piece, concrete proof of a six-month journey from your first program to a robot.
Share it, in both communities
Show your TARS in the six-month graduation showcase to mark the completion of the whole program. And do one more thing: take it back to the TARS-AI community. You built their robot; sharing your build, your signature behavior, and what you learned in their Discord is how open source works, it inspires the next builder, it thanks the people whose work you stood on, and it might teach the community something new. Giving back to the project that gave you the design closes the loop in the best possible way.
TARS exists because a community of makers built and shared it. Post your finished robot, your custom behaviors, and your lessons learned in their Discord, and consider contributing improvements back to the project. That is how the whole thing keeps getting better for the next person who reaches this month:
GitHub: github.com/TARS-AI-Community/TARS-AI · Discord: discord.gg/AmE2Gv9EUt · Docs: docs-tars-ai.vercel.app
Look back at the arc. You learned to program and to solder and design your own board. You ran AI on a Raspberry Pi and gave it sight, speech, and a mind. You automated a home and learned to secure it. And you finished by 3D printing and assembling a robot that combines all of it, standing on the shoulders of an open-source community you are now part of. TARS is not a separate project; it is the sum of everything you learned, and the proof that you went from your first circuit to an AI-powered inventor. Congratulations.