[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-70474":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":10,"languages":9,"totalLinesOfCode":9,"stars":11,"forks":12,"watchers":13,"openIssues":13,"contributorsCount":14,"subscribersCount":14,"size":14,"stars1d":13,"stars7d":13,"stars30d":15,"stars90d":14,"forks30d":14,"starsTrendScore":16,"compositeScore":16,"rankGlobal":9,"rankLanguage":9,"license":17,"archived":18,"fork":18,"defaultBranch":19,"hasWiki":20,"hasPages":20,"topics":21,"createdAt":9,"pushedAt":9,"updatedAt":32,"readmeContent":33,"aiSummary":34,"trendingCount":14,"starSnapshotCount":14,"syncStatus":35,"lastSyncTime":36,"discoverSource":37},70474,"ogcode","prasenjeet-symon\u002Fogcode","prasenjeet-symon","The browser-native coding agent that runs entirely on your machine. Plans features, executes in parallel across git branches, and remembers everything. One binary. Zero cloud dependencies.",null,"Go",122,9,1,0,42,3,"MIT License",false,"main",true,[22,23,24,25,26,27,28,29,30,31],"ai","ai-agent","automation","code-generation","coding-assistant","developer-tools","go","golang","llm","pair-programming","2026-06-12 02:02:34","# Ogcode\n\n[![Discord](https:\u002F\u002Fimg.shields.io\u002Fdiscord\u002F1373677337985056828?label=Discord&logo=discord&logoColor=white&color=5865F2)](https:\u002F\u002Fdiscord.gg\u002FJQP9t8y2Zv)\n\n![Ogcode Demo](assets\u002Fogcode_intro.gif)\n\nAn agentic coding assistant with a web UI, written in Go.\n\nOgcode acts as a pair programmer that actually codes with you. It doesn't just suggest snippets; it understands your entire codebase, plans complex features, and executes them by creating branches and PRs automatically. It can even run multiple tasks in parallel across different branches, allowing you to ship entire features in a fraction of the time. Whether you're hunting a bug or building a zero-to-one feature, Ogcode handles the heavy lifting so you can focus on the architecture.\n\nOgcode gives you two ways to work with AI on your codebase:\n\n- **Build Mode** — Chat directly with an AI agent that can read, write, edit, and execute code in your project.\n- **Plan Mode** — Collaboratively plan a feature or refactor with a read-only planning agent, then break the plan into tasks. Each task gets its own git branch and an isolated agent session. Completed tasks auto-create pull requests. Multiple tasks can run in parallel, drastically speeding up complex feature implementations.\n\n---\n\n## Roadmap\n\n- [ ] **Advanced Task Planning & Parallel Execution**: Enhanced plan decomposition, allowing users to manually or automatically assign coding agents to tasks, with full parallel execution support.\n- [ ] **AI Daily Standups**: Voice-enabled daily meetings where all agents assigned to a project can report progress and discuss their work. This includes natural voice interactions, project-specific memory, and the ability to take and remember human feedback.\n- [ ] **Ogland Integration**: A dedicated space within Ogcode to connect external services like Slack, Email, Jira, and more, enabling their use during the planning phase.\n- [ ] **Agentic Deployment**: Full end-to-end agentic deployment support for all major cloud providers, starting with AWS.\n\n---\n\n## Agentic Session Memory 🧠\n\n![Agentic Memory Demo](assets\u002Fagentic_memory.gif)\n\nOgcode's **Agentic Session Memory** revolutionizes how AI coding assistants handle context in long-running sessions. Instead of sending the entire conversation history to the LLM (which quickly becomes expensive and hits token limits), Ogcode intelligently extracts, stores, and retrieves only the relevant context needed for each query.\n\n### Key Benefits\n\n| Feature | Impact |\n|---------|--------|\n| **~70% Token Savings** | Drastically reduced API costs on long sessions |\n| **Infinite Context** | No practical limit on session length or codebase size |\n| **Higher Accuracy** | Only relevant memories are retrieved per query |\n\n### Enable Agentic Memory\n\n```bash\nexport OGCODE_AGENTIC_MEMORY_MODE=true\n```\n\n### Token Savings Example\n\n| Session Length | Traditional | With Agentic Memory | Savings |\n|----------------|-------------|---------------------|---------|\n| 50 messages | ~25K tokens | ~8K tokens | **68%** |\n| 200 messages | ~100K tokens | ~28K tokens | **72%** |\n| 1000 messages | ~500K tokens | ~120K tokens | **76%** |\n\n*Actual savings vary based on codebase complexity and conversation patterns.*\n\n---\n\n## Installation\n\n### macOS \u002F Linux\n\n**Via Homebrew:**\n\n```bash\nbrew tap prasenjeet-symon\u002Fogcode\nbrew install ogcode\n```\n\n**Via curl:**\n\n```bash\ncurl -fsSL http:\u002F\u002Fogcode.xyz\u002Finstall.sh | sh\n```\n\nThe installer auto-detects your platform, downloads the latest release, and installs to `\u002Fusr\u002Flocal\u002Fbin` (uses `sudo` if needed).\n\n### Windows\n\n```powershell\nirm http:\u002F\u002Fogcode.xyz\u002Finstall.ps1 | iex\n```\n\nThis downloads the latest release, extracts it to `%LOCALAPPDATA%\\ogcode`, and adds it to your PATH automatically.\n\n**Via winget (after next release):**\n\n```powershell\nwinget install prasenjeet-symon.ogcode\n```\n\n**Manual install:**\n\n1. Go to the [releases page](https:\u002F\u002Fgithub.com\u002Fprasenjeet-symon\u002Fogcode\u002Freleases) and download `ogcode_Windows_x86_64.zip` (or `_arm64.zip` if you have an ARM device).\n2. Extract the zip file to a folder (e.g. `C:\\Tools\\ogcode`).\n3. Add that folder to your `Path` environment variable:\n   - Press `Win + S`, search for **Edit environment variables for your account**\n   - Under **User variables**, find `Path` and click **Edit**\n   - Click **New** and add the path to your ogcode folder (e.g. `C:\\Tools\\ogcode`)\n   - Click **OK** on all dialogs\n4. Open a new PowerShell or Command Prompt and run:\n\n```powershell\nogcode version\n```\n\n### Go Install\n\n```bash\ngo install github.com\u002Fprasenjeet-symon\u002Fogcode@latest\n```\n\n### Docker\n\n```bash\ndocker run -p 8080:8080 -v $(pwd):\u002Fworkspace -w \u002Fworkspace ghcr.io\u002Fprasenjeet-symon\u002Fogcode:latest\n```\n\n---\n\n## Configuration\n\nOgcode auto-detects available AI providers based on environment variables. No config files are required.\n\n### Required: AI Provider\n\nSet at least one API key (or use Ollama):\n\n| Variable | Provider |\n|----------|----------|\n| `ANTHROPIC_API_KEY` | Anthropic (Claude) |\n| `OPENAI_API_KEY` | OpenAI (GPT) |\n| `OPENROUTER_API_KEY` | OpenRouter |\n| `OLLAMA_API_KEY` | Ollama Cloud (see below) |\n| `OLLAMA_BASE_URL` | Ollama (local \u002F cloud URL) |\n\n#### Ollama (local models)\n\nOgcode auto-detects Ollama on macOS\u002FLinux if the binary is installed at a common path. On Windows, or if you have a non-standard install, set `OLLAMA_BASE_URL` explicitly.\n\n**Local setup (default):**\n\n```bash\n# macOS \u002F Linux — auto-detected if ollama is installed\nollama serve\nogcode\n\n# Or be explicit on any OS:\nexport OLLAMA_BASE_URL=http:\u002F\u002Flocalhost:11434\u002Fv1\nogcode\n```\n\nOn Windows (PowerShell):\n\n```powershell\n$env:OLLAMA_BASE_URL = \"http:\u002F\u002Flocalhost:11434\u002Fv1\"\nogcode\n```\n\n**Remote or Ollama Cloud:**\n\n```bash\nexport OLLAMA_BASE_URL=https:\u002F\u002Fapi.ollama.com\u002Fv1   # or your custom endpoint\nexport OLLAMA_API_KEY=your-api-key                  # required for cloud \u002F authenticated endpoints\nogcode\n```\n\nOn Windows (PowerShell):\n\n```powershell\n$env:OLLAMA_BASE_URL = \"https:\u002F\u002Fapi.ollama.com\u002Fv1\"\n$env:OLLAMA_API_KEY = \"your-api-key\"\nogcode\n```\n\n**Set a default model:**\n\n```bash\nexport OLLAMA_MODEL=codellama   # defaults to qwen3-coder-next if not set\n```\n\nAvailable models in the UI include: `qwen3`, `codellama`, `llama3.1`, `deepseek-coder-v2`, `mistral`, and others. Any model you have pulled in Ollama will work — just select it from the model dropdown in the web UI.\n\n### Optional: Agentic Memory\n\nTo give the agent long-term memory across sessions, set:\n\n```bash\nexport OGCODE_AGENTIC_MEMORY_MODE=true\n```\n\nThis connects to an MCP-compatible memory server (configure via `MCP_SERVER_*` env vars).\n\n### Optional: Custom Models\n\nYou can add custom models (e.g. fine-tuned endpoints) through the web UI at **Settings → Models**.\n\n---\n\n## Usage\n\n### Start in Build Mode (default)\n\n```bash\nogcode\n```\n\nOpens the web UI at `http:\u002F\u002Flocalhost:8080`. Chat with the agent, ask it to read files, write code, run commands, or search the codebase.\n\n### Start in Plan Mode\n\n```bash\nogcode plan\n```\n\nOpens the planning interface. Describe what you want to build. The planning agent will understand your codebase and discuss the approach with you. When you are satisfied, click **Lock Plan** — the agent breaks it into tasks with dependencies, effort, and complexity estimates.\n\n### Use a custom port\n\n```bash\nogcode -p 3000\nogcode plan -p 3000\n```\n\n### Check version\n\n```bash\nogcode version\n```\n\n---\n\n## The Plan Mode Workflow\n\n1. **Describe** — Open a new plan and describe your goal. The planning agent reads your codebase and helps refine the approach.\n2. **Lock** — When ready, lock the plan. The agent generates a structured task breakdown.\n3. **Review** — View tasks in the Kanban board. Each task has a title, description, effort (S\u002FM\u002FL\u002FXL), complexity, and dependencies.\n4. **Execute** — Start tasks. Each one gets its own git branch and an isolated agent session. You can watch the agent work in real time.\n5. **Complete** — When a task finishes, the agent commits its changes and a pull request is automatically created. Ogcode manages dependencies between tasks and runs independent tasks in parallel.\n6. **Retry** — If a task fails, retry it. The stale branch is removed and the task starts fresh.\n\nPlans are archived as markdown files in `.ogcode\u002Farchives\u002F` once all tasks are completed.\n\n---\n\n## Community\n\nJoin the Ogcode community on Discord for discussions, support, and updates:\n\n[![Join us on Discord](https:\u002F\u002Fimg.shields.io\u002Fdiscord\u002F1373677337985056828?label=Join%20Discord&logo=discord&logoColor=white&color=5865F2&style=for-the-badge)](https:\u002F\u002Fdiscord.gg\u002FJQP9t8y2Zv)\n\n- Ask questions and get help\n- Share feedback and feature ideas\n- Stay up to date with releases and announcements\n\n---\n\n## License\n\nMIT License — see [LICENSE](LICENSE) for details.\n","Ogcode 是一个基于 Web 的代码开发助手，使用 Go 语言编写，能够处理从理解代码库到规划特性、创建独立分支、编写代码并提交 PR 的整个开发流程。其核心功能包括通过 AI 代理与开发者协同工作，不仅提供代码片段建议，还能自动理解和执行复杂的代码任务，并支持多任务并行处理。项目特别适合需要快速迭代和复杂特性开发的场景，如新功能构建或大规模重构。此外，Ogcode 利用智能会话记忆技术，在长会话中有效减少 API 成本的同时保持高准确性，为开发者提供了高效且经济的编码辅助解决方案。",2,"2026-06-11 03:32:30","CREATED_QUERY"]