The Silent Pact

In a world where sound attracts deadly shadow creatures, players are guardians of a silent sanctuary. They must work together to solve puzzles, gather resources, and defend their home—all without making a peep. Communication happens through a built-in visual sign language system, performed via controller gestures or webcam tracking. The game’s narrative unfolds through animated vignettes and environmental storytelling, centered on Deaf culture and resilience. Haptic feedback vibrates through the controller to convey events like footsteps or alarms, ensuring no one misses critical information. Multiplayer is seamless: up to four players sign to each other, with no voice chat or audio cues. The ultimate goal is to rebuild the sanctuary and teach new members the silent way of life.

Mainstream games overwhelmingly rely on audio for immersion, communication, and gameplay cues, directly excluding Deaf and hard-of-hearing players. Even subtitles or visual indicators often fall short in multiplayer coordination. This game scratches the itch for a truly inclusive experience where audio is not a crutch but a deliberate design constraint—players must become fluent in a new visual language to succeed. It’s not just accessible; it’s built from the ground up around the strengths of the Deaf community: powerful visual communication and heightened awareness of non-audio stimuli.

• Develop an intuitive, learnable in-game sign language with ~200 gestures for common commands, items, and emotions.
• Design cooperative puzzles that require non-verbal communication (e.g., one player sees a pattern, another must interpret and replicate).
• Ensure 100% audio-free gameplay by balancing visual and haptic cues for all essential feedback.
• Create a story that authentically represents Deaf culture, with character arcs centered on identity and community.
• Achieve sub-100ms gesture recognition latency via webcam or controller input.
• Host a closed beta with at least 50 Deaf players to validate accessibility and fun.

• Not a competitive multiplayer game; focus on cooperation and shared goals.
• No voice chat or text-to-speech features; the sign language is the only communication channel.
• Not a real sign language simulator; the in-game vocabulary is simplified and gamified for learning.
• No procedural audio generation; all sound feedback is replaced by haptics or visuals.
• Not a roguelike or permadeath; campaigns are story-driven and session-based.

• Game Engine: Unity (for real-time 3D and broad platform support)
• Gesture Recognition: Unity Asset Store 'Gesture Framework' for controller-based input; OpenCV + MediaPipe for webcam-based hand tracking (optional)
• Haptics: Custom haptic profiles using Unity's Input System for Xbox/PS5 controllers
• Multiplayer: Azure PlayFab for matchmaking, Photon Quantum for deterministic state sync
• Art: Blender for 3D assets, Aseprite for 2D UI icons (sign language dictionary)
• Project Management: Notion, Git, Discord for community playtests

The game runs on a client-server architecture. Each player's device captures gesture input (controller buttons/joystick combos or webcam hand positions) and runs a local gesture classifier. Recognized gestures are sent as lightweight action packets (e.g., 'player_1_sign_help') to the authoritative game server. The server processes these actions, updates the game state, and broadcasts relevant events to all clients. Haptic cues are triggered client-side based on the event type (e.g., a 'danger' event triggers a pulse pattern). The sign language system is a lookup table mapping gesture sequences to in-game actions and visual icons displayed above each player's avatar. The camera is fixed isometric, ensuring all players can see each other's signs at all times.

• Gesture recognition accuracy: false positives could break puzzles; requires extensive tuning and player calibration.
• Learning curve: non-Deaf players may struggle to adopt a new language, leading to frustration or abandonment.
• Puzzle balance: puzzles must be solvable without audio but not trivial; need iterative testing with Deaf playtesters.
• Multiplayer latency: gesture transmission lag could disrupt real-time cooperation; need rollback netcode or deterministic lockstep.
• Accessibility conflicts: webcam tracking may exclude players without cameras; must default to controller input.

• Should the sign language have a dictionary that pauses gameplay to teach new signs, or should it be learned organically through contextual hints?
• How to handle emotes / quick communication (e.g., 'follow me') without slowing down the puzzle flow?
• What haptic patterns are intuitive for danger, success, or item pickup? Should we rely on existing industry standards?
• Are there legal considerations with using real ASL signs? Should the in-game language be purely invented to avoid cultural appropriation?
• Can we implement a 'sign check' system where players can practice and get feedback on their gestures?

The core team of three part-time developers spent a year on a prototype but hit a wall with gesture recognition latency on average webcams. Simultaneously, the lead designer moved to a different city and the project lost momentum. It remains a beloved blueprint, waiting for a more robust tech foundation or a dedicated team.

The project originally included a 'soundscape' mode for hearing players to experience the silence, but it was removed to keep the focus on Deaf-first design. All playtest sessions would be conducted in sign language with Deaf moderators.