Ward/en is a single-player strategy game on PC that blends programming with salvage operations in derelict spacecraft. Players write code directives to control autonomous drones that explore wrecks, retrieve cargo, and return data without any live steering from the operator.
Gameplay
The core loop centers on writing directives in the WARD/EN language, a small imperative programming language that includes functions, enums, for loops, callbacks, modules, and pure functions. Hardware interactions occur through a limited set of calls that each consume a tick and return results. All higher-level behaviors, such as route planning in move_to, exist as editable library code that players can inspect, modify, or replace entirely.
After composing a directive as a project of source files, players dispatch it to freeze the code and send a drone into a wreck known only through a recovered schematic. The drone operates independently, with its battery serving as the time limit amid patrolling threats. Upon completion or failure, players receive the recovered cargo, a mission report, and an updated map of the explored area. Drones that exhaust their power remain lost inside the wreck.
Advanced features support multi-drone crews that communicate via fitted transceivers, allowing scouts to relay findings to haulers. Callbacks enable step-by-step decision making during movement, such as responding to contacts or threats. Integrated tooling provides code completion, hover information, go-to-definition, rename, and diagnostics that account for the drone's specific equipment fit, available both in the in-game Workbench and through a VS Code extension with language server support.
Debugging occurs on a test bench that executes directives against prebuilt scenarios. Players can set breakpoints, step through code at various levels, inspect call stacks and local variables, and monitor drone position, energy, and cargo. Resumed paused runs produce identical results to uninterrupted executions.
Game Modes
Training Grounds offers eight distinct problem types that a procedural generator uses to construct wrecks: Unknown Hulk, Infested Hulk, Salvage Run, Search & Recover, Hazardous Hulk, Signal Hunt, Multi-Drone Operation, and Anything Goes. Each run operates under a single budget constraint with the full equipment catalogue available from the start. Up to four drones can operate simultaneously, each with its own fit and directive, with total costs measured against the budget.
Scenarios are fully reproducible through a shareable code consisting of the problem type and a seed value. Every objective remains reachable by at least one valid equipment configuration within battery limits. Runs track multiple metrics including spend, drones lost, time taken, salvage value recovered, and instructions executed, each compared against personal bests for that specific scenario code. Different solutions optimize for different priorities such as cost, speed, or safety.
Equipment and Systems
The catalogue contains sixty-eight parts across eight hull types and thirteen module sockets. Higher-cost components provide capabilities unavailable on cheaper alternatives, creating trade-offs between hardware investment and code sophistication. Sensing options include spot scanners, forward lidar, acoustic mappers that penetrate walls, and seismic sensors. Action tools encompass cutters, breaching charges, security bypass kits, grapples, and relays. Survival modules feature decoys, smoke, jammers, stunners, and specialized hulls. Navigation can rely on purchased co-processors or custom route planners written by the player. Power systems center on batteries that limit runtime, while transceivers range from basic pagers to laser links, supported by telemetry buffers for outage recovery.
Signal limitations mean the ship receives only uplink packets, with corridors or obstacles able to interrupt transmission. The simulation maintains private state for creatures, geometry, and intent, so players work exclusively from the data returned in each packet.
Learning and Presentation
A guided tutorial covers the full workflow. A seventeen-lesson language course uses pinned expeditions on the training wreck, while a five-lesson Fast Track condenses core concepts for experienced programmers. Training runs support playback speeds up to eight times normal. The interface renders through a CRT-style compositor with phosphor persistence, bloom, scanlines, and grain, alongside an optional high-contrast mode for extended sessions.
Is It Worth Playing?
Ward/en targets players who enjoy programming challenges combined with strategic planning and resource management in a single-player format. The absence of a campaign or unlock progression keeps focus on iterative improvement of directives across reproducible scenarios. With the game listed as coming soon and no user reviews available yet, suitability depends on interest in writing and debugging code to solve salvage problems rather than traditional action or narrative elements. Those drawn to custom scripting, debugging tools, and open-ended optimization across multiple performance metrics will find the systems align directly with that preference.