![](\"https://img.shields.io/badge/Python-3.10%2B-3776AB?style=for-the-badge&logo=python&logoColor=white\"){kind=icon}
![](\"https://img.shields.io/badge/CrewAI-Multi--Agent-FF6B35?style=for-the-badge&logo=robot&logoColor=white\"){kind=icon}
![](\"https://img.shields.io/badge/Streamlit-1.35%2B-FF4B4B?style=for-the-badge&logo=streamlit&logoColor=white\"){kind=icon}
\n \n \n ![](\"https://img.shields.io/badge/OpenAI-GPT--4o--mini-412991?style=for-the-badge&logo=openai&logoColor=white\"){kind=icon}
\n ![](\"https://img.shields.io/badge/Mode-Demo%20%7C%20Live-22C55E?style=for-the-badge\"){kind=icon}
\n ![](\"https://img.shields.io/badge/License-MIT-16A34A?style=for-the-badge\"){kind=icon}
\n \n
\n A role-specific AI mock interview simulator built with CrewAI and Streamlit.
\n Pick a job role, answer 6 curated questions, and get instant AI-powered feedback โ all in a polished dark UI.\n
\n Works fully without an API key. Add one to unlock live GPT-4o-mini feedback via CrewAI.\n
\n\n---\n\n## ๐ Table of Contents\n\n- [Overview](#-overview)\n- [Demo vs Live Mode](#-demo-vs-live-mode)\n- [UI Screens](#-ui-screens)\n- [Features](#-features)\n- [Supported Roles](#-supported-roles)\n- [Workflow](#-workflow)\n- [Project Structure](#-project-structure)\n- [Getting Started](#-getting-started)\n- [How It Works](#-how-it-works)\n- [CrewAI Agent](#-crewai-agent)\n- [Tech Stack](#-tech-stack)\n- [License](#-license)\n\n---\n\n## ๐ง Overview\n\nThis project extends a basic CrewAI mock interviewer into a **role-specific** interview simulator. It supports four job roles โ Data Scientist, Web Developer, Product Manager, and UI/UX Designer โ each with 6 hand-crafted questions.\n\nThe app auto-detects whether an OpenAI API key is present and switches between Demo Mode (rule-based feedback, zero API calls) and Live Mode (CrewAI Feedback Coach agent powered by GPT-4o-mini). Mentors running the app with a valid key get full AI feedback automatically.\n\n---\n\n## ๐ Demo vs Live Mode\n\n| | ๐ก Demo Mode | ๐ข Live Mode |\n|---|---|---|\n| API key required | No | Yes โ `OPENAI_API_KEY` in `.env` |\n| Feedback engine | Length-aware rule-based logic | CrewAI agent โ GPT-4o-mini |\n| All UI features | โ Full | โ Full |\n| API calls made | โ None | โ Per answer |\n| Sidebar badge | ๐ก Demo Mode | ๐ข Live Mode |\n\nThe mode is detected at startup โ no code changes needed to switch between them.\n\n---\n\n## ๐ฅ๏ธ UI Screens\n\n### Screen 1 โ Landing & Role Selection\n- Centered hero title and subtitle\n- 2ร2 interactive role card grid โ each card shows an icon, role name, and skill tag\n- Selected card highlights with a role-specific color glow border\n- Sidebar displays \"How it works\" steps and the list of available roles\n- Confirm selection and start with a single button click\n\n### Screen 2 โ Interview\n- Role icon + name header\n- Gradient progress bar (`#6C63FF โ #00B4D8`) tracking question completion\n- Styled question card with a role-color left border accent\n- Answer textarea with a **live word-count indicator** โ turns green in the ideal 30โ120 word range\n- Collapsible \"Previous answers\" panel showing all completed Q&A with feedback\n- Sidebar updates live with role, answered count, and % progress\n- Sidebar restart button available at any point\n\n### Screen 3 โ Summary\n- Gradient completion banner (green) with role and question count\n- 3-column metrics strip: total questions ยท average answer length ยท role\n- Full expandable session review โ every question with your answer and a styled blue AI feedback card\n- \"Try Another Role\" button to restart cleanly\n\n---\n\n## โจ Features\n\n- ๐จ Dark theme โ custom CSS, Inter font, `#0f1117` base, gradient accents throughout\n- ๐ Interactive role cards โ color-coded per role with hover lift and selection glow\n- ๐ Live word-count guide โ real-time feedback on answer length as you type\n- ๐ Live sidebar stats โ role, answered count, progress % update after each submission\n- ๐ฌ Styled feedback cards โ blue gradient panel visually separates AI feedback from answers\n- ๐ Collapsible history โ previous Q&A stays accessible without cluttering the active question\n- ๐ Summary metrics โ question count, average word count, role displayed as st.metric cards\n- ๐ Restart anywhere โ from sidebar during interview or from the summary screen\n- ๐ก/๐ข Mode badge โ always visible in sidebar so it's clear which mode is active\n\n---\n\n## ๐ผ Supported Roles\n\n| Role | Icon | Color | Focus Areas |\n|------|------|-------|-------------|\n| Data Scientist | ๐ง | `#6C63FF` | ML ยท Statistics ยท Modelling |\n| Web Developer | ๐ป | `#00B4D8` | APIs ยท Frontend ยท HTTP |\n| Product Manager | ๐ | `#F77F00` | Strategy ยท Metrics ยท Roadmap |\n| UI/UX Designer | ๐จ | `#E63946` | Design ยท Research ยท Accessibility |\n\n---\n\n## ๐ Workflow\n\n```mermaid\nflowchart TD\n A([๐ User Opens App]) --> B{OPENAI_API_KEY\\npresent in .env?}\n B -- Yes --> C[๐ข Live Mode\\nCrewAI + GPT-4o-mini]\n B -- No --> D[๐ก Demo Mode\\nRule-based feedback]\n C --> E[Landing โ Role Selection Cards]\n D --> E\n\n E --> F[User selects a role\\nand clicks Start Interview]\n F --> G[Load 6 role-specific questions]\n\n G --> H{Questions\\nremaining?}\n H -- Yes --> I[Show styled question card\\nwith gradient progress bar]\n I --> J[User types answer\\nLive word-count indicator]\n J --> K[Submit Answer]\n K --> L{Mode?}\n L -- Live --> M[CrewAI Feedback Coach\\nAgent โ GPT-4o-mini]\n L -- Demo --> N[Length-aware\\nrule-based feedback]\n M --> O[Render styled feedback card\\nSave to session history]\n N --> O\n O --> H\n\n H -- No --> P[Summary Screen\\n3-col metrics + full Q&A review]\n P --> Q{Try another role?}\n Q -- Yes --> E\n Q -- No --> R([โ Session complete])\n\n style A fill:#238636,color:#fff\n style R fill:#238636,color:#fff\n style M fill:#6C63FF,color:#fff\n style N fill:#F77F00,color:#fff\n style C fill:#1a7f37,color:#fff\n style D fill:#9e6a03,color:#fff\n style P fill:#0d2137,color:#58a6ff\n```\n\n> Full `.mmd` source: [`Flow/workflow.mmd`](Flow/workflow.mmd)\n\n---\n\n## ๐ Project Structure\n\n```\ncrewai-role-based-ai-interviewer/\nโ\nโโโ app.py # All-in-one Streamlit app\nโ # ยท Role config & questions\nโ # ยท Demo / Live feedback logic\nโ # ยท Custom CSS theme\nโ # ยท Landing, Interview, Summary screens\nโ\nโโโ requirements.txt # Python dependencies\nโโโ .env.example # API key template โ copy to .env\nโโโ .gitignore # Ignores .env, __pycache__, .venv, etc.\nโโโ LICENSE # MIT License\nโโโ README.md # This file\nโ\nโโโ Flow/\n โโโ workflow.mmd # Mermaid source for the workflow diagram\n```\n\n---\n\n## ๐ Getting Started\n\n### Prerequisites\n\n- Python 3.10 or higher\n- _(Optional)_ An [OpenAI API key](https://platform.openai.com/api-keys) for Live Mode\n\n### Installation\n\n**1. Clone the repo**\n\n```bash\ngit clone https://github.com/SANJAI-s0/crewai-role-based-ai-interviewer.git\ncd crewai-role-based-ai-interviewer\n```\n\n**2. Create a virtual environment**\n\n```bash\npython -m venv .venv\n\n# Windows\n.venv\\Scripts\\activate\n\n# macOS / Linux\nsource .venv/bin/activate\n```\n\n**3. Install dependencies**\n\n```bash\npip install -r requirements.txt\n```\n\n**4. Set up your API key** _(skip this for Demo Mode)_\n\n```bash\ncp .env.example .env\n```\n\nEdit `.env`:\n```env\nOPENAI_API_KEY=sk-...your-key-here...\n```\n\n**5. Run the app**\n\n```bash\nstreamlit run app.py\n```\n\nOpens at `http://localhost:8501`.\n\n---\n\n## โ๏ธ How It Works\n\n**1. Mode detection**\nOn startup, the app reads `OPENAI_API_KEY` from `.env`. If a valid key is found, `LIVE_MODE = True` and a ๐ข badge appears in the sidebar. Otherwise, ๐ก Demo Mode is used โ no API calls are ever made.\n\n**2. Role selection**\nThe landing page renders 4 role cards in a 2ร2 grid. Clicking a card highlights it with a role-specific color glow. The selected role and its skill tags are confirmed below before starting.\n\n**3. Interview loop**\nQuestions are shown one at a time. A gradient progress bar tracks completion. As the user types, a live word-count indicator turns green when the answer is in the ideal 30โ120 word range.\n\n**4. Feedback**\n- _Live Mode_: `get_live_feedback()` lazily imports CrewAI and LangChain, builds a `Feedback Coach` agent, runs a single-task Crew, and returns 2-3 sentences of constructive feedback.\n- _Demo Mode_: `get_demo_feedback()` checks word count and returns a contextual response from a curated pool โ no imports, no API calls.\n\n**5. Summary**\nAfter all 6 questions, a summary screen shows a completion banner, a 3-column metrics strip, and a full expandable review of every Q&A with styled feedback cards.\n\n---\n\n## ๐ค CrewAI Agent\n\nUsed only in Live Mode. Instantiated lazily โ never created when running in Demo Mode.\n\n```python\nAgent(\n role=\"Interview Feedback Coach\",\n goal=\"Provide concise, constructive feedback on interview answers\",\n backstory=(\n \"You are an experienced hiring manager and interview coach. \"\n \"You give short, honest, and encouraging feedback. Keep it to 2-3 sentences.\"\n ),\n llm=ChatOpenAI(model=\"gpt-4o-mini\", temperature=0.7),\n)\n```\n\nEach submitted answer creates a `Task` for this agent and runs it through a single-agent `Crew`. The result is rendered in a styled blue feedback card.\n\n---\n\n## ๐ ๏ธ Tech Stack\n\n| Technology | Version | Purpose |\n|------------|---------|---------|\n| [Streamlit](https://streamlit.io/) | โฅ 1.35 | Web UI, session state, custom CSS |\n| [CrewAI](https://github.com/joaomdmoura/crewAI) | โฅ 0.28 | Multi-agent orchestration (Live Mode) |\n| [LangChain OpenAI](https://python.langchain.com/) | โฅ 0.1 | LLM wrapper for GPT-4o-mini |\n| [OpenAI](https://platform.openai.com/) | โฅ 1.30 | GPT-4o-mini API (Live Mode) |\n| [python-dotenv](https://pypi.org/project/python-dotenv/) | โฅ 1.0 | `.env` file loading |\n\n---\n\n## ๐ License\n\nThis project is licensed under the [MIT License](LICENSE).\n","readmeExcerpt":"๐๏ธ CrewAI Role-Based AI Interviewer B{OPENAI_API_KEY\\npresent in .env?}\n B -- Yes --> C[๐ข Live Mode\\nCrewAI + GPT-4o-mini]\n B -- No --> D[๐ก Demo Mode\\nRule-based feedback]\n C --> E[Landing โ Role Selection Cards]\n D --> E\n\n E --> F[User selects a role\\nand clicks Start Interview]\n F --> G[Load 6 role-specific questions]\n\n G --> H{Questions\\nremaining?}\n H -- Yes --> I[Show styled question card\\nwith gradient progress bar]\n I --> J[User types answer\\nLive word-count indicator]\n J --> K[Submit Answer]\n K --> L{Mode?}\n L -- Live --> M[CrewAI Feedback Coach\\nAgent โ GPT-4o-mini]\n L -- Demo --> N[Length-aware\\nrule-based feedback]\n M --> O[Render styled feedback card\\nSave to session history]\n N --> O\n O --> H\n\n H -- No --> P[Summary Screen\\n3-col metrics + full Q&A review]\n P --> Q{Try another role?}\n Q -- Yes --> E\n Q -- No --> R([โ
Session complete])\n\n style A fill:#238636,color:#fff\n style R fill:#238636,color:#fff\n style M fill:#6C63FF,color:#fff\n style N fill:#F77F00,color:#fff\n style C fill:#1a7f37,color:#fff\n style D fill:#9e6a03,color:#fff\n style P fill:#0d2137,color:#58a6ff"},{"language":"text","snippet":"crewai-role-based-ai-interviewer/\nโ\nโโโ app.py # All-in-one Streamlit app\nโ # ยท Role config & questions\nโ # ยท Demo / Live feedback logic\nโ # ยท Custom CSS theme\nโ # ยท Landing, Interview, Summary screens\nโ\nโโโ requirements.txt # Python dependencies\nโโโ .env.example # API key template โ copy to .env\nโโโ .gitignore # Ignores .env, __pycache__, .venv, etc.\nโโโ LICENSE # MIT License\nโโโ README.md # This file\nโ\nโโโ Flow/\n โโโ workflow.mmd # Mermaid source for the workflow diagram"},{"language":"bash","snippet":"git clone https://github.com/SANJAI-s0/crewai-role-based-ai-interviewer.git\ncd crewai-role-based-ai-interviewer"},{"language":"bash","snippet":"python -m venv .venv\n\n# Windows\n.venv\\Scripts\\activate\n\n# macOS / Linux\nsource .venv/bin/activate"},{"language":"bash","snippet":"pip install -r requirements.txt"},{"language":"bash","snippet":"cp .env.example .env"}],"parameters":null,"dependencies":[],"permissions":[],"extractedFiles":[],"languages":["python"],"docsSourceLabel":"GITHUB OPENCLEW","editorialOverview":"Role-specific AI mock interviewer built with CrewAI and Streamlit. Features a dark UI with interactive role cards, live word-count feedback, sidebar session stats, and a full summary screen. Runs in Demo Mode (no API key) or Live Mode (GPT-4o-mini via CrewAI) โ auto-detected from .env. ๐๏ธ CrewAI Role-Based AI Interviewer