December 2025 Devlog
Base Mods and Bloom Mechanics
Game Design Pivot
The PodRun spec underwent a major pivot in late November. After implementing pod mechanics with proper physics, playtesting with Janel revealed that bumping into units and pushing them around was the most engaging aspect. She suggested a soccer-like objective—bringing a ball to an edge or designated area.
This feedback crystallized a new vision. The previous spec—evolving into more complex creatures—felt isolated and lacked team-focused mechanics for a multiplayer environment. The new design centers on bases that players feed, upgrade, protect, and expand. Extraction rewards scale with team performance, making it an optionally cooperative experience.
Team Formation
Since PodRun must remain a sandbox loop-free environment, traditional team structures don't apply. Players form tribes organically through a merge trigger: hold the merge key near other pods or a base unit. After cooldown, nearby paired pods (up to 3) differentiate into a base unit, opening the bloom menu.
Base Blooming
Once players accumulate enough seed mass and merge, they enter the bloom stage. The viewport locks to the base, and players select from 2 randomly drawn upgrades (from a pool of 8). Teams vote collectively; the majority choice wins. After ~10 seconds, the base unlocks and players can spawn as mods.
Base Mods
Mods are base extensions with specialized functions. They cost seed and pulp mass to spawn, consume pulp mass for energy and abilities, and can be mounted and controlled by players. Unlike pods, mods remain bound to their base except during special abilities.
Current implementations:
- Bay — Similar to pod mouth, with push (spacebar) and mass eject (Q)
- Tail — In development
Core Loop
Feed the base enough seed mass to reach max level, then extract. Extraction destroys the base core (without decomposition) and rewards gear: abilities, pets, skins.
Engine Optimizations
Collision System Rewrite
Rewrote broadphase detection and blob collision algorithms. Results:
- 2–4× performance improvement
- Simplified, more accurate collision resolution
- Reduced cognitive overhead and cleaner API
- Target: ~100 users with 2–4k semi-sparse units on current server ($15 dedicated, 8× 2.4GHz cores)
View Processing
Implemented delta views, reducing network/processing overhead ~4× (1 Mbit → 200 Kbit for units alone). Field and view data not yet delta-compressed.