{"id":13456034,"url":"https://github.com/jw782cn/RepoChat-200k","last_synced_at":"2025-03-24T09:31:18.357Z","repository":{"id":228970999,"uuid":"772525195","full_name":"jw782cn/RepoChat-200k","owner":"jw782cn","description":"⚡Chat with GitHub Repo Using 200k context window of Claude instead of RAG!⚡","archived":false,"fork":false,"pushed_at":"2024-03-30T07:18:48.000Z","size":3923,"stargazers_count":164,"open_issues_count":3,"forks_count":26,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-03-14T20:13:09.971Z","etag":null,"topics":["chatbot","claude","llm","openai","openrouter","streamlit"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/jw782cn.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null}},"created_at":"2024-03-15T11:12:12.000Z","updated_at":"2025-02-14T08:50:31.000Z","dependencies_parsed_at":"2024-04-04T06:15:13.865Z","dependency_job_id":null,"html_url":"https://github.com/jw782cn/RepoChat-200k","commit_stats":null,"previous_names":["jw782cn/repochat-200k"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jw782cn%2FRepoChat-200k","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jw782cn%2FRepoChat-200k/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jw782cn%2FRepoChat-200k/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/jw782cn%2FRepoChat-200k/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/jw782cn","download_url":"https://codeload.github.com/jw782cn/RepoChat-200k/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":245243252,"owners_count":20583592,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["chatbot","claude","llm","openai","openrouter","streamlit"],"created_at":"2024-07-31T08:01:15.201Z","updated_at":"2025-03-24T09:31:18.345Z","avatar_url":"https://github.com/jw782cn.png","language":"Python","funding_links":[],"categories":["Python"],"sub_categories":[],"readme":"# RepoChat in 200k - No RAG!\n\n⚡Chat with GitHub Repo Using 200k context window of Claude instead of RAG!⚡\n\n\u003cimg src=\"examples/example.gif\" height=\"400\"\u003e\n\n\n## Why RepoChat - No RAG?\n\nTake the advantage of Claude 200k! Put all examples and codes to the contexts!\n\n**We need copilot rather than agent sometimes!**\n\n### 1. Not just Chat, But Write Codes with the Lastest Documents\n\nHave troubles memorizing all the apis in llama-index or langchain? \n\nNo worries, just include the components examples and the documents from the repo and let Claude Opus - the strongest model and long context window of 200k to write your agent for you!\n\n### 2. No copy paste, Just Select\n\n\u003cimg src=\"examples/repos.png\" height=\"300\"\u003e\n\nDownload/Clone your Repo from Github then just select the files you'd like, I got you covered on constructing the prompt.\n\n### 3. No RAG, You decide\n\nI've seen many Chat with Repo projects, they all have the painpoints:\n\n`Which files do this query need?`\n\nThey use **embedding search** in Code database but most of the time I already knew what documents I'm refering... So make your own choices each time when you are coding.\n\n- Coding Frontend? Just select components and examples.\n\n- Coding Agents? Just select Jupyter Notebook of langgraph.\n\n- Coding RAG? Just select Jupyter Notebook of llamaindex.\n\n\u003cimg src=\"examples/selection.png\" height=\"400\"\u003e\n\n\n### 4. More Use Cases!\n\n#### Use llamaindex doc\n\nselect llamaindex example of pipeline to write RAG graph.\n\n\u003cimg src=\"examples/llamaindex.png\" height=\"400\"\u003e\n\n#### Use Shadcn doc to write frontend\n\nselect examples and components definition.\n\n\u003cimg src=\"examples/shadcn.png\" height=\"400\"\u003e\n\n\n## Suggestions!\n\n1. You can use Haiku for most of the case.\n2. Change models based on tasks.\n3. Change files based on tasks.\n4. Clone Repos You like! \n5. Usually I will include README of repo to help Claude Understand better\n6. USE `COUNT TOKENS` on the sidebar to see how many tokens you will send!!!\n\n\u003cimg src=\"examples/count.png\" height=\"200\"\u003e\n\n## Features\n\n1. **Repository Download**: Users can provide a GitHub repository URL, and the application will automatically download and analyze the repository.\n2. **File and Folder Selection**: Users can select specific files or folders from the repository to include in the LLM's input.\n3. **Language Filtering**: Users can filter the files by programming language to focus the LLM's understanding on specific parts of the codebase.\n4. **Token Limit**: Users can set a token limit to control the amount of information sent to the LLM, which can be useful for performance or cost considerations.\n5. **Chat Interface**: Users can interact with the LLM through a chat-style interface, allowing them to ask questions or request code generation based on the repository contents.\n6. **Streaming Output**: The LLM's responses are displayed in a streaming fashion, providing a more engaging and real-time user experience.\n\nCurrently I only supported Openrouter. Planing to add more and refactor someday.\n\n## Get Started\n\n1. **Environment Settings**: Run `pip install -r requirements.txt` to set up environment.\n\n2. **Create a .env file**: Create a `.env` file in the root directory of the project and add your OpenRouter API key (Recommended):\n```bash\nOPENROUTER_API_KEY=your_openrouter_api_key_here\n```\nI recommend [OpenRouter](https://openrouter.ai/) because it has all models!\n\nIf you want to use OpenAI GPT models, add your `openai api key` as well.\n\n```bash\nOPENAI_API_KEY=your_openai_api_key_here\n```\n\n3. **Run the application**: Run the `app.py` script using Streamlit:\n```bash\nstreamlit run app.py\n```\n4. **Use the application**: Follow the instructions in the application to download a GitHub repository, select files and folders, and chat with the LLM.\n\nIf you encounter some issues with repo, you can always delete the repo dir in ./repos dir and download it again.\n\n## Configuration\n\nThe application's behavior can be customized through the following configuration options:\n\n- **Model**: The specific LLM model to use (e.g., \"anthropic/claude-3-haiku\", \"anthropic/claude-3-opus\").\n- **Temperature**: The temperature parameter that controls the \"creativity\" of the LLM's responses.\n- **System Prompt**: The initial prompt given to the LLM to set the desired behavior.\n\nThese settings can be adjusted in the sidebar of the Streamlit application.\n\n## Contributing\n\nIf you'd like to contribute to the RepoChat-200k project, please feel free to submit issues or pull requests on the [GitHub repository](https://github.com/jw782cn/RepoChat-200k).\n\n## License\n\nThis project is licensed under the [MIT License](LICENSE).","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjw782cn%2FRepoChat-200k","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fjw782cn%2FRepoChat-200k","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fjw782cn%2FRepoChat-200k/lists"}