___ ___ _ _ _ _ _ | \/ | | | (_) | | | | | | | . . | __ _ ___| |__ _ _ __ ___ | | | | ___ ___| | __ | |\/| |/ _` |/ __| '_ \| | '_ \ / _ \ | |/\| |/ _ \/ _ \ |/ / | | | | (_| | (__| | | | | | | | __/ \ /\ / __/ __/ < \_| |_/\__,_|\___|_| |_|_|_| |_|\___| \/ \/ \___|\___|_|\_\
For full documentation, please refer to our site designated to the Ouja Board
For Machine Week, our architecture group wanted something that felt playful and theatrical, but still demanded real machine integration: motion control, hidden mechanisms, tolerances, and a clean user experience. After some chaotic ideation (robotic tufter, cocktail mixer, etc.), we landed on the Ouija board: a “haunted” spirit board that appears to answer questions on its own… while hiding the entire mechanism inside a custom-built table. Its legacy was rooted in the mechanical turks of the 18th century.
The user interacts with a square Ouija board set into a designed table. You ask a question, place your hands near the planchette, and the cursor glides toward letters/numbers as if guided by an unseen presence. The intent was “effortless haunted object”: the surface looks clean and graphic, the motion feels smooth, and the engineering disappears.
We originally imagined additional stage effects (like candles that tip over on certain questions), but we kept refining scope as the machine came together.
Avi organized us into sub-teams and assigned captains to coordinate dependencies. I ended up straddling multiple teams because the board had to interface cleanly with both the table + gantry + cursor/end effector. I started as “Board lead,” but quickly expanded into integration work:
Early team brainstorm: quick votes, fast scope decisions, lots of interface dependencies.
The end effector needed to pull the planchette (cursor) from below the board using magnetic coupling. The core questions were:
A Ouija board is basically theatrical graphic design pretending to be a board game. The mood needed to be dialed all the way up: ornate, occult, dramatic — but still refined enough to feel intentional rather than parody. To set the baseline, we ordered the cheapest “Spirit Board” Amazon could ship and used it as a reference for scale, atmosphere, and kitsch factor.
Reference board: fast calibration for scale, vibe, and layout.
Our board needed to be perfectly square to integrate with the custom table. To iterate quickly, we used ChatGPT to generate initial motifs and compositions and then refined them toward our needs. Full prompt conversation: Ouija Board Concept Prompts (ChatGPT share).
We also had a hard constraint: the gantry could not reach all the way to the edges. That meant text + motifs needed to be offset inward so the planchette could reliably hit targets.
Early Rhino composition: square format + reachable layout constraints.
We wanted extra layers of imagery — celestial icons, runes, skulls, filigree — but also wanted to avoid accidentally using loaded or culturally insensitive symbols. (Researching occult symbolism online is… a lot of double-checking.)
We discovered a very practical limitation: AI is notoriously bad at rendering clean alphabets (missing letters, duplicated letters, no “A”, etc.). So after generating style direction, we moved into Illustrator to rebuild the layout into editable, manufacturable vectors.
This part was shockingly annoying (and absolutely the real work):
AI doesn’t generate manufacturable geometry — so in practice we rebuilt the board manually to make it laser-safe.
The board needed a slick top surface so the planchette would glide with minimal friction. That meant etching from the back, not the front. By divine intervention, Chris gifted us a sheet of 1/4" plexiglass (thanks, Chris).
We tested:
Our finalists were white graphics on black, or black graphics on white. We ultimately intended to go white-on-black… but I was exhausted, mixed up the order, and produced the inverse. It still looked legitimately spooky.
After a long laser job, we produced the final board:
For carpentry, we consulted the oracle of N51: Jen. With limited time (and limited appetite for fancy joinery), we chose hidden screws and reliable construction.
Because the table had to house a gantry being developed in parallel, we didn’t have a stable reference for a while. So we locked a strategy:
A tree recently died behind N52 and has been quietly furnishing the shop with free lumber. We continued the tradition: rough band-saw cuts, planed boards to consistent thickness, then cut/assembled/sanded the table for gantry integration.
Candle holder concept I sketched out.
Mounting day was pure HTMAA energy: throw out over-engineering and do what works. We iteratively stacked and fastened scrapwood blocking until the extrusions landed at the right height and alignment. We cross-checked our as-builts against Rhino/Fusion models and adjusted until motion cleared the board and the portal.
Once the board, cursor/end effector, and table interfaces converged, the whole thing started to feel “effortless” — which was the goal. The most satisfying part was seeing the magic effect emerge from mundane realities: tolerances, friction, spacing, magnet coupling, and mounting.