PlanToCodeTechnical System Walkthrough
A technical walk-through of how PlanToCode is built so you can study the architecture, reuse the patterns, or build a similar planning pipeline. The desktop app relies on external LLM providers configured with your API keys.
System Walkthrough Video
A short, end-to-end walkthrough of the planning pipeline: task input, file discovery jobs, plan generation, merge instructions, and execution handoff.
- Tauri shell and React UI surface the workflow
- Rust background jobs stream progress updates
- SQLite stores sessions, jobs, and logs
- LLM plans are generated and merged with structure
System Architecture and Implementation Walkthrough
Each panel below shows the desktop app and backend collaborating: Tauri shell, Rust job processors, SQLite persistence, and model orchestration coordinating planning and execution.
System Walkthrough Video
A short, end-to-end walkthrough of the planning pipeline: task input, file discovery jobs, plan generation, merge instructions, and execution handoff.
- ✓Tauri shell and React UI surface the workflow
- ✓Rust background jobs stream progress updates
- ✓SQLite stores sessions, jobs, and logs
- ✓LLM plans are generated and merged with structure
File Discovery Pipeline
A four-stage Rust workflow: LLM-assisted root selection, regex filtering, relevance scoring, and extended path finding to build a focused file set.
- ✓Root folder selection uses the directory tree and task prompt
- ✓Regex filter generates pattern groups and applies git ls-files
- ✓Relevance scoring chunks file contents with token estimates
- ✓Extended path finder expands context with file + tree data
Persisted File Context
Each workflow stage writes results to background_jobs so file sets can be reused across sessions and inspected later.
- ✓Workflow stages stored as job records
- ✓Selected file lists persisted as JSON responses
- ✓Session included_files reused across jobs
- ✓SQLite history survives restarts
Video Analysis Ingestion
Screen recordings are sent to the `/api/llm/video/analyze` endpoint with prompts and FPS settings to generate structured summaries.
- ✓Multipart upload includes duration_ms and framerate
- ✓Model format is provider/model (google/* required)
- ✓Usage and cost tracked in server billing logs
- ✓Summary stored as job metadata artifacts
Multi-Model Plan Generation
ImplementationPlanProcessor streams plan drafts from full file contents; merge jobs consolidate multiple drafts into one plan.
- ✓Plan jobs include selected file contents + directory tree
- ✓Structured plan metadata captured per job
- ✓Merge prompt uses <source_plans> and <user_instructions>
- ✓Final plan stored alongside source drafts
Background Job Monitoring
Rust job processors stream progress and state transitions to the UI while persisting job history in SQLite.
- ✓Created, queued, preparing, running, completed/failed/canceled
- ✓Streaming updates via Tauri events
- ✓Token usage captured per run
- ✓Cancel long-running jobs
Prompt and Model Configuration
Runtime model settings are fetched from `/api/config/desktop-runtime-config`; prompt overrides are stored in SQLite.
- ✓Per-task allowed models and defaults
- ✓System prompts served by the server API
- ✓Project-level prompt overrides in project_system_prompts
- ✓Local key_value_store for runtime preferences
Workflow Automation Buttons
Copy buttons insert templated prompts with task context for handoff into terminals or external tools.
- ✓Templates sourced from task model config
- ✓Placeholders resolved against the active plan
- ✓Handoff to PTY sessions or clipboard
- ✓Actions tied to job metadata for audit
Plan Merge Instructions
ImplementationPlanMergeProcessor merges multiple plan drafts using XML-tagged source plans and optional instructions.
- ✓Source plans pulled by job ID
- ✓Merge instructions stored in metadata
- ✓File contents + directory tree add context
- ✓Merged plan stored alongside inputs
Usage and Cost Ledger
Server-side usage entries and job metadata capture model usage across providers.
- ✓Per-job token and cost metadata
- ✓Provider-aware usage entries
- ✓Billing endpoints expose usage summaries
- ✓Audit trail for model spend
Explore the architecture in context
Map each UI panel to the underlying processors, then dive deeper into the architecture docs and build guides.