Getting Started
Learn how to use ConvoVerse's powerful AI features in minutes.
Start a New Chat
Chat with any LLM provider
Fork Conversations
Branch off and explore alternatives
Response Tiers
3 response tiers for any task
Knowledge Base
Ask questions about your documents
Start a New ChatCore Feature
Chat with any LLM provider from a single interface. Switch models mid-conversation, choose response tiers, and leverage tools.
How to Start
- 1Click "New Chat" in the sidebar or dashboard
- 2Choose your LLM provider (OpenAI, Anthropic, Google, xAI, Groq, Cohere, DeepSeek, Mistral, RunPod, MiMo)
- 3Select a Response Tier (Quick, Balanced, or Deep)
- 4Enable tools if needed (Web Search, Knowledge Base, Calculator, File Read, Code Execution)
- 5Start chatting!
Supported Providers
Switch Models Anytime
You can change models mid-conversation. Start with a powerful model for complex reasoning, then switch to a faster/cheaper model for follow-ups. Your conversation history is preserved.
Fork Conversations
Branch off from any point in a conversation to explore alternative directions without losing the original.
Why Fork?
- Explore alternatives - Try different prompts from the same starting point
- Compare models - See how different LLMs respond to the same question
- Build on others' work - Fork public conversations and continue them
- Preserve attribution - Original author is always credited
How to Fork
- 1Open any conversation (yours or public)
- 2Click the Fork button (or hover on a message and fork from there)
- 3Choose your LLM and continue the conversation
- 4Your fork is saved separately with a link to the original
Response Tiers
Choose the right reasoning strategy for your task - from quick answers to deep analysis with verification.
Quick
Fast answers, lowest cost
Balanced
Plans before answering
Deep
Thorough analysis, cross-verified
Tip: Start with Quick for simple questions. Use Balanced when you need step-by-step reasoning. Deep is best for complex tasks that benefit from thorough analysis and cross-verification.
Knowledge Base & Tools
Extend AI capabilities with document Q&A and integrated tools.
Knowledge Base (RAG)
- My Files — organize documents in collections; scope which collections the AI searches in chat
- Upload PDF, Word, text, CSV, JSON, HTML, YAML, and Markdown files — with Google Document AI OCR for scanned documents and embedded images
- Ask questions and get answers with source citations
- Hybrid search (BM25 + vector) and neural reranking for relevance
Available Tools
Import Existing Conversations
Already have valuable conversations in ChatGPT, Claude, or Grok? Import them to continue or share.
Browser Extension (Recommended)
- 1Install from Chrome Web Store
- 2Visit any ChatGPT, Claude, or Grok conversation
- 3Click the extension icon to import
Why Import?
- Continue with any model - Pick up where you left off with a different LLM
- Unified search - Search across all your conversations in one place
- Share & collaborate - Make conversations public for others to fork
- Never lose a conversation - Preserve valuable AI interactions
Search & Discover
Your Conversations
Full-text and semantic search across all your chats. Find any conversation instantly.
Public Feed
Discover valuable conversations shared by the community. Fork and learn from others.
Knowledge Base
Search your uploaded documents with AI-powered retrieval.
Quick Tips
Getting Started
- • Start with a New Chat to explore the interface
- • Try different Response Tiers to see the difference
- • Use DeepSeek models for affordable, high-quality chat
- • Fork public conversations to learn from others
Pro Tips
- • Switch models mid-conversation for different perspectives
- • Use Deep tier for complex research with verification
- • Enable Web Search for real-time information
- • Upload documents to the Knowledge Base for AI Q&A
Troubleshooting
Chat not responding
Check your credits balance and try again
Response taking too long
Try Quick tier for faster responses
Extension not working
Make sure you're logged in to ConvoVerse
Knowledge base not finding docs
Try rephrasing your question or check document status