README.md

metadata

title: Tell Me
emoji: 💬🌿
colorFrom: indigo
colorTo: green
sdk: streamlit
app_file: fresh_app_v2.py
pinned: false
tags:
  - streamlit
short_description: Mental wellbeing chat (research)

🌿 Tell Me — A Mental Well-Being Space

Tell Me is a safe space for individuals seeking some well-being advice or a self-reflection space. It also provides the research community to simulate some LLM generated client-therapist synthetic data. This is a research prototype, not a medical device.

Key Components of Tell Me:***

• Tell Me Assistant

  Tell Me Assistant is a Mental Well-being Chatbot designed to help individuals process their thoughts and emotions in a supportive way. It is not a substitute for professional care, but it offers a safe space for conversation and self-reflection. The Assistant is created with care, recognizing that people may turn to it during moments of initial support. Its goal is to make such therapeutic-style interactions more accessible and approachable for everyone.

  fresh_app_v2.py interconnected with rag.py and llm_models.py to power the Assistant with context using RAG

• Simulate a Conversation
  This generates a synthetic client–therapist conversation from a short client profile. It helps create sample data for research and lets professionals inspect the dialogue quality. Outputs are created by an LLM and can guide future fine-tuning or evaluation. Multi‑turn, role‑locked Therapist ↔ Client dialogue from a brief persona (see llm_models.py).

• Well‑being Planner (CrewAI)

  1. Transcript analysis (themes, emotions, triggers)
  2. 7‑day plan (CBT/behavioral steps, routines, sleep hygiene, social micro‑actions)
  3. Guided meditation script + MP3 (gTTS/Edge/Coqui/ElevenLabs)
     Implemented in crew_ai.py, surfaced in the Planner tab in fresh_app_v2.py.

• Evaluation
  Use prep_responses.py and judge.py to prepare and score generations using LLM as a judge and also the results of conducted Human Evaluation; see Results/ for artifacts (e.g., gpt4o/5 eval).

---

Repository Structure

.
├─ Results/                 # Evaluation outputs / artifacts (e.g., gpt4o eval)
├─ index_storage/           # Vector index built by rag.py
├─ rag_data/                # Source docs for RAG
├─ src/                     # Streamlit template seed
├─ bg.jpg                   # App background
├─ config.toml              # Streamlit config (dark mode default, etc.)
├─ crew_ai.py               # CrewAI pipeline (planner + meditation TTS)
├─ fresh_app_v2.py          # Main Streamlit app
├─ judge.py                 # Evaluation judge
├─ llm_models.py            # Prompt builders + simulate‑conversation helpers
├─ prep_responses.py        # Prep helper for evaluation
├─ rag.py                   # Simple RAG indexing/query helpers
├─ requirements.txt         # Python dependencies
├─ Dockerfile               # Optional container build
├─ .gitattributes
└─ README.md                # You are here :)

---

Quickstart

1) Python setup

# Python 3.10+ recommended
python -m venv .venv
source .venv/bin/activate           # Windows: .venv\Scripts\activate
pip install -r requirements.txt

2) Environment variables

Create a .env in the project root (same folder as fresh_app_v2.py). Minimal example:

# Optional TTS configuration for the guided meditation
# TTS_PROVIDER=gtts               # or: edge | coqui | elevenlabs
# ELEVEN_API_KEY=...              # if using ElevenLabs
# EDGE_VOICE=en-US-JennyNeural    # if using edge-tts
# COQUI_MODEL=tts_models/multilingual/multi-dataset/xtts_v2

Some tabs may allow choosing models/keys in the UI.
The Planner currently works with the key above (and/or an in‑tab field if present in your build).

3) Run the app

streamlit run fresh_app_v2.py

Open the URL Streamlit prints (usually http://localhost:8501).

---

Using the App

UI View Recommendation

Note: On the streamlit app please ensure that you have selected the Dark Mode in settings to get the best UI Experience of the App

Simulate a Conversation 🧪🤖

1. In that tab, paste a Client Profile (e.g., Age 24 student; recently moved... sleep irregular...).
2. Click Generate Synthetic Dialogue to produce a multi‑turn conversation.
3. Optionally Download Transcript.

Well‑being Planner 📅🧘

1. Ensure you provide open ai api key for running this module (i.e paste a key in the tab if the field is available).
2. Upload one .txt transcript (client–therapist chat).
3. Click Create Plan & Meditation.
4. The app displays:
   • Transcript Summary
   • 7‑Day Well‑being Plan (Markdown, Day 1 … Day 7)
   • Meditation Script and an MP3 player
     (audio saved locally as guided_meditation.mp3)

RAG (optional)

• Place your files into rag_data/.

• Build/update the index (if needed):

  python rag.py
  

• Use the app’s RAG controls to query your docs (index artifacts stored in index_storage/).

Evaluation (optional)

• Use prep_responses.py to format generations and judge.py to score them.
• Outputs/examples are kept under Results/.

---

Streamlit Configuration

• config.toml sets app defaults (e.g., dark mode). Example:

[theme]
base = "dark"

Adjust as needed per Streamlit docs.

---

Docker (optional)

# Build
docker build -t tellme-assistant .

# Run (exposes Streamlit on 8501)
docker run --rm -p 8501:8501 --env-file .env tellme-assistant

---

Troubleshooting

• AuthenticationError / “You didn’t provide an API key.”
  Ensure .env includes open_ai_key_for_crew_ai=sk-... (or provide the key in‑tab if available) and restart Streamlit so the new env is loaded.

• Only meditation shows but not the plan
  Update to the latest crew_ai.py that collects and returns summary, plan, and meditation, and ensure the tab renders all three fields.

• TTS provider errors
  Install the provider’s dependency (pip install edge-tts, pip install TTS, pip install elevenlabs) and set the related env vars.

• Ollama (if used in other tabs)
  Start the daemon and pull a model:

  ollama serve
  ollama pull llama3.1:8b-instruct
  

---

Tech Stack

• UI: Streamlit
• LLMs: OpenAI (planner), plus optional Anthropic/Ollama in other tabs
• Agents: CrewAI (via LiteLLM under the hood)
• RAG: Simple local index (rag.py, index_storage/)
• TTS: gTTS / Edge‑TTS / Coqui TTS / ElevenLabs (configurable)

---

Roadmap

• In‑tab API key entry for the CrewAI planner (UI‑first flow)
• Configurable model/provider for planner
• Save generated plans/MP3s into Results/ with timestamped filenames

---

License

MIT

---

Acknowledgments

• Streamlit template seed
• CrewAI & LiteLLM ecosystem
• TTS libraries: gTTS, Edge‑TTS, Coqui TTS, ElevenLabs

Acknowledgment of AI Assistance

Some parts of this project code was generated or refined with the assistance of GPT-5. All outputs were reviewed, tested, and integrated by the authors.