CodebuddyTask #1Hagicode
Mainstream Agents Supported10x Parallel Management EfficiencyOpenSpec Reduces Hallucinations
Task / Agent
IdeateProposeExecuteArchive
CodebuddyTask #1KiroTask #2
DeepAgentsTask #3
Gemini CLITask #4Claude CodeTask #5Completed Tasks0
Live Efficiency1.8xrelative to a single serial worker
People Are Using It
Community keeps growing, thanks to every contributor
Three Core Features
Redefine your AI coding experience
SMART
Smart
OpenSpec workflow, boost AI coding efficiency
300%
Efficiency Boost
A complete 9-stage proposal workflow from idea to archive, with AI assistance throughout. Boost your coding efficiency by 3x.
IdeaStart from idea
ProposalAI generates proposal
ReviewAuto review
TasksTask breakdown
CodeSmart coding
TestAuto testing
RefactorCode refactor
DocsDocs generation
ArchiveKnowledge archive
EFFICIENT
Efficient
Multi-agent, multi-instance execution that keeps your quota moving
Traditional single-agent / single-instance
→
Multi-agent parallel matrix
Throughput Boost10x+ parallel
Current supported Agent CLIs such as Claude Code, Codex, GitHub Copilot, OpenCode, and Hermes can work across multiple instances at the same time, so proposals, fixes, and reviews move in parallel instead of waiting in a serial queue.
Parallel execution matrixMultiple agents / multiple instances
11 currently supported Agent CLIs
Claude CodeCodexGitHub CopilotOpenCodeHermesQoderCLIKiroKimiGeminiDeepAgentsCodebuddy
Claude Code
Proposal / review
Draft specCheck designArchive review
Codex
Implementation / fixes
Build feature ABuild feature BRegression fix
Qwen · GLM
Analysis / support
Trace issuePolish copyCapture insights
Multiple agents are driving multiple instances right now
INTERESTING
Fun
Hero Dungeon makes campus collaboration and practice feel alive
Hero Dungeons
Turn generation, execution, and prompt workbenches into dungeon runs
Captain Roster
Assign the right hero captain to each dungeon
Battle Reports
Review XP, levels, and execution results after each push
Fun no longer depends on achievement badges. Hero Dungeon, captain rosters, prompt visuals, and battle reports turn daily AI coding practice into a workflow that feels more like a coordinated adventure for campus labs and student teams.
Today's dungeon reportHero Battle
4Dungeons pushed
Lv.27Captain level
12.8MXP today
When several dungeons move in parallel, the report rolls up experience, levels, and execution outcomes into one readout.
Captain ready
Proposal DungeonOpenSpec generate / apply / archive
2 instances active
AutoTask DungeonParallel fixes and batch automation
Visual sync ready
Prompt DungeonPrompt preview and style workbench
Current captain roster
Spec Strategist
Proposal captain
Patch Runner
Execution captain
Prompt Artist
Visual support
Product Showcase
See the current HagiCode interface through its multi-agent workflow, Hero Dungeon workspace, and battle report views
Light Theme Main Interface
A bright workspace that keeps project context and task flow easy to scan.
Dark Theme Main Interface
A dark workspace designed for long-running multi-agent sessions and status tracking.
Real-time token usage
Track token movement per session so parallel workloads stay visible and controllable.
Multi-agent workspace
Different agents can drive multiple instances at the same time for drafting, fixing, and reviewing.
Hero Dungeon roster
Organize captains and members by dungeon to turn workflow control into a visual workspace.
Hero Battle report
Review experience, level progression, and execution output in one battle-style summary.
Real Product Walkthroughs
See Hagicode in real coding sessions
Start with the core daily workflow, then jump into a playful AI side quest and a GPT Codex validation run inside Hagicode.
Featured walkthrough
Daily Hagi Half-Hour: AI Multi-Task Programming Practice
A guided overview of Hagicode's everyday workflow, agent orchestration, and the core experience new visitors should understand first.
Open on BilibiliFocused demo
AI Playing Games While Coding
A playful clip that shows how the assistant can stay lively and surprising during real coding sessions instead of feeling like a dry code generator.
Open on BilibiliFocused demo
GPT Codex in Hagicode: Live Trial
A focused validation run of GPT Codex inside Hagicode so prospective users can judge how the model performs in the actual product.
Open on BilibiliChoose Your Installation Method
Choose the most suitable installation method based on your usage scenario