Theadwise

Threadwise: Multidimensional AI Chatbox The Problem Standard chat interfaces force a linear timeline on nonlinear thinking. Trying a different approach usually means editing a past prompt (which destroys subsequent context) or dragging the AI through a messy history of discarded ideas, losing context and forcing re-explaining. The Solution Threadwise introduces Branching Conversations to LLM interactions. Inspired by Git, every prompt and response is a node in a directed graph, letting you fork the conversation from any historical point. A backend LineageAssembler walks the ancestor chain of your active branch, intelligently excluding siblings and future branches. The LLM only sees the exact chronological timeline leading to that specific fork. Key Features & Capabilities True Branching Mechanics (Working): Fork mid-thread. Context is strictly scoped to the ancestor chain, preventing cross-branch leakage. Multi-Provider BYOK Routing (Working): Bring Your Own Key for Google AI Studio (Gemini) and Fireworks AI. Switch models on the fly while keeping context lineage intact. Lineage-Aware Context Loading (Working): Smart history assembly. Create stateless temporary branches so dead-end thoughts never pollute your main timeline. Summary Nodes (Planned): Compact long branches into summary nodes, cutting token usage while preserving semantic context. Eagle View Graph (Planned): A visual graph mapping all branches, forks, and nodes to navigate complex project topologies. Product Vision Threadwise transforms AI interactions into a durable, multidimensional workspace for power users. By breaking free from linear chat constraints, no idea is lost, and every tangent can be safely explored.