Search papers, labs, and topics across Lattice.
This paper introduces AgentRoom, a real-time collaborative editing protocol designed for concurrent coding among multiple agents, leveraging Conflict-free Replicated Data Types (CRDTs) to manage a shared filesystem. The system was evaluated using five frontier coding-CLI models on various backend coding tasks, revealing that AgentRoom significantly reduces task abandonment rates compared to traditional solo approaches and exhibits less variability in performance. The findings suggest that effective coordination among agents, rather than mere parallelism, is crucial for optimizing multi-agent coding efficiency.
AgentRoom slashes task abandonment rates in multi-agent coding by emphasizing coordination over mere parallel execution.
Concurrent multi-agent coding promises division of labor across modules, robustness through redundancy, and parallel exploration at the natural granularity of multi-file projects. Realtime collaborative editing protocols solve this coordination problem for human teams via Conflict-free Replicated Data Types (CRDTs), but the LLMs underneath generate one token at a time and existing multi-agent coding systems inherit this serial limit: they either sequence agents through phase handoffs or pool independent samples without coordination, and a single agent abandons up to half of hard tasks with a one-file stub-and-exit. AgentRoom is a realtime collaborative editing protocol for concurrent coding agents. Its runtime layer exposes file-level claim, status, and broadcast as MCP tools on a CRDT-merged shared filesystem. Five frontier coding-CLI models ran four backend coding tasks, with cross-language checks in Python DevBench and Rust+axum. For CLI-stable models, AgentRoom with 2 agents abandons fewer tasks than Solo and has less run-to-run variation. At matched-compute, one positive mean LLM-judge contrast puts AgentRoom over parallel-merge. The other contrast, a bundle probe, puts full AgentRoom above each partial case: an ordering rather than a percentage split. Coordination, not parallelism or CRDT-merge, bears the load.
