Documentation

New Conversation

Click 'New Chat' in the sidebar or navigate to the home page. Each conversation maintains its own context and history.

Send Messages

Type in the input area and press Enter to send. Use Shift+Enter for line breaks. Messages support Markdown, code blocks, and LaTeX math.

File Attachments

Attach PDFs, images (PNG, JPG, GIF, WebP), and code files. Click the attachment button or drag files into the input area. Multiple files supported per message.

Conversation History

All conversations are saved automatically. Access them from the 'Chats' page in the sidebar. Search, rename, or delete conversations at any time.

Quick Switching

Click any tab to instantly switch between active conversations. The tab bar shows chat names for easy identification.

Create & Close

Click the + button to open a new chat tab. Close any tab with its X button. Your remaining tabs stay intact.

Persistent Tabs

Tabs are saved to your session and restored when you reload the page. Pick up exactly where you left off.

Live Streaming

Responses stream in real-time via Server-Sent Events. Watch the AI generate text token by token.

Pause / Resume / Stop

Control active streams with pause, resume, and stop buttons that appear during generation. Stop immediately cancels the response.

Text-to-Speech

Play any AI response as audio using the TTS button on messages. Toggle auto-TTS from the input menu to hear all responses automatically.

Copy & Screenshot

Copy any message to clipboard with one click. Use the screenshot feature to capture message exchanges as shareable images with gradient backgrounds.

Regenerate

Click the retry button (refresh icon) below any AI response to regenerate it. The AI produces a fresh answer using the same conversation context.

+ Context

Use the '+ context' option to add additional instructions before retrying. Guide the AI toward a better answer by specifying what to change, expand, or focus on.

Normal

Default fast response mode. Best for simple questions and quick interactions.

Thinking

Extended analysis mode. The AI takes more time to consider its response, showing its thinking process in a collapsible block with elapsed time.

Reasoning

Step-by-step logical reasoning. Ideal for complex problems, math, and code debugging. Shows detailed reasoning chain.

Team of Experts

Multi-perspective analysis. The AI consults multiple specialized viewpoints before synthesizing a comprehensive answer.

Built-in Personalities

Choose from Default, Efficient, Friendly, or Professional modes. Each adjusts the AI's communication style to match your preference.

Custom Personality Plugins

Install personality plugins from the marketplace to add entirely new AI behaviors. Installed personalities appear in the input menu alongside built-in options.

Personalization Settings

Set your display name and personal preferences in Settings → Personalization. The AI uses these to tailor every response to your style.

Markdown Rendering

Full Markdown support: bold, italic, lists, tables, blockquotes, and links. All rendered beautifully in the chat interface.

Syntax-Highlighted Code

Code blocks with automatic language detection and syntax highlighting. Each block includes a copy button for quick extraction.

LaTeX Math

Write mathematical expressions using LaTeX syntax. Inline math with $ and display math with $$. Rendered with KaTeX.

Highlights & Bookmarks

Bookmark important AI responses with the highlight button. Access all bookmarks from the Highlights tab on the Chats page. Search, review, and jump back to any highlighted message.

Visual Indicator

A ghost icon appears when incognito is active, so you always know your privacy status.

Session-scoped

Data persists during your browser session but is cleared when you close the tab. Nothing appears in your chat history.

Connect to Ollama

Open the Connect dropdown in the input toolbar and click 'Connect to Ollama'. PolyChat automatically detects available models running on localhost:11434.

Model Selection

Once connected, select any available Ollama model from the detected list. Switch between Ollama and cloud models at any time.

Works Everywhere

Ollama connection is available both on the main page and inside active chat sessions. Your selected model persists across messages.

Requirements

Ollama must be running on your local machine (localhost:11434). Install Ollama and pull models before connecting.

Preset Templates

Choose from ready-made team configurations: Product Development, Code Review, Debate & Refine, Creative Writing, and more. Each template sets up a team of agents with complementary roles.

Custom Teams

Build your own agent teams from scratch. Add multiple agents with custom roles, instructions, and characters. Enable or disable individual agents on the fly.

Sequential Responses

Agents respond in order. The first agent replies to your message, then subsequent agents can build on or respond to previous agents' answers.

Agent-to-Agent Replies

Agents can target specific agents to reply to, creating natural multi-perspective discussions.

Persistent Config

Group chat configurations are saved per conversation and automatically restored when you return.

Split View

Split your chat into multiple panels — horizontally or vertically. Each panel can hold a different conversation. Max 2 levels deep to keep things manageable.

Artifacts

When the AI generates code, analysis, or reports, they're automatically extracted into artifacts displayed in a side panel. Three types: Code, Reason, and Report.

Version History

Every artifact edit creates a new version. Browse the full version timeline and revert to any previous state.

Playgrounds

Interactive playground links let you view and edit generated content in a dedicated panel. Enable Playgrounds in Settings → Personalization.

Chat History

All conversations appear on the Chats page, sorted by most recent activity. See message counts, last updated times, and group chat badges at a glance.

Search

Search across all conversations by name or content. Filter results to quickly find what you need.

Rename & Delete

Rename any conversation for better organization. Delete individual chats or bulk-delete all at once (with confirmation).

Project Assignment

Assign conversations to projects for organized, context-sharing workflows. Link chats to projects directly from the Chats page.

Token Tracking

View tokens used and remaining in your current billing cycle. Breakdown shows input tokens (your messages) and output tokens (AI responses) separately.

Conversation Limits

Track your daily conversation count against your plan's limit. See how many conversations you have left today.

Reset Timing

See exactly when your usage counters reset. Daily limits reset every 24 hours; token allowances reset with your billing cycle.