ask_bubble_ai 1.5.1

A premium floating AI chat bubble widget for Flutter. Easily embed a chatbot assistant powered by Gemini, OpenAI, Claude, DeepSeek, and OpenRouter with custom styling, streaming, and navigation.

1.5.1

• Interactive Copy Button Feedback: Added dynamic checkmark icon (Icons.check_rounded) feedback when clicking the inline copy button on chat bubbles. Reverts back to copy icon after 2 seconds.
• Enabled copy button for both user and assistant chat bubbles.
• Comply with pub.dev limits by restricting topics list to 5 items in pubspec.yaml.

1.5.0

• OpenRouter provider support: Added AskAiProvider.openRouter as a fifth AI provider option. OpenRouter is a unified API gateway that gives access to hundreds of models from OpenAI, Anthropic, Meta, Mistral, Google, and more through a single OpenAI-compatible endpoint (openrouter.ai/api/v1/chat/completions) with Bearer token authentication.
• Includes recommended HTTP-Referer and X-Title headers as per OpenRouter API guidelines.
• Default model for OpenRouter must be specified via modelName (e.g. openai/gpt-4o, anthropic/claude-sonnet-4-5, meta-llama/llama-3.3-70b-instruct).
• Updated pubspec.yaml description, version, and topics to include openrouter.
• Updated README.md with OpenRouter API key instructions, provider table, and configuration property descriptions.

1.4.1

• Minor fixes and corrections.

1.4.0

• DeepSeek provider support: Added AskAiProvider.deepSeek as a fourth AI provider option. Uses DeepSeek's OpenAI-compatible Chat Completion API (api.deepseek.com) with Bearer token authentication.
• Default model for DeepSeek is deepseek-v4-flash (overridable via modelName).
• Updated pubspec.yaml description and topics to include deepseek.
• Updated README.md with DeepSeek API key instructions, provider table, and configuration property descriptions.
• Updated example app to demonstrate the DeepSeek provider option.

1.3.9

• Issues fixed and README.md updated.

1.3.8

• Some corrections made.

1.3.7

• Some issues fixed and README.md updated.

1.3.6

• Fixed markdown formatting in README.md to ensure the License section is displayed as a separate section instead of inside a code block.

1.3.5

• Updated screenshot paths in README.md to use direct hosted image URLs, resolving display issues on pub.dev.
• Simplified markdown parser logic in AskAiPanel by removing a redundant ternary expression.
• Set the default provider to Gemini in the example application to align with the default environment key and model configuration.

1.3.4

• Fixed pub.dev image rendering compatibility by replacing absolute raw GitHub URLs (which fail on private repositories) with a standard Markdown table using relative local package paths (assets/).

1.3.3

• Fixed README screenshots not displaying on pub.dev due to private/moved repository URL paths by changing image paths to local package-relative directories (assets/).

1.3.2

• Added high-quality screenshots to README.md to showcase the customizable UI of the floating chat bubble and assistant panel.
• Secured example app API key using String.fromEnvironment and configured a .env template local configuration (with .gitignore rules).

1.3.1

• Added avatarWidget: Widget? to AskAiHeaderStyle. When set, it fully replaces the default icon inside the circular avatar container, allowing any widget (e.g. Image.asset, Image.network, a logo, or a Stack combining an icon + badge) to be used as the assistant avatar. Falls back to avatarIcon when not provided.

1.3.0

• Gemini streaming responses: Gemini provider now streams tokens in real-time for a live-typing effect inside the chat panel.
• Inline markdown rendering: AI responses now render **bold**, *italic*, and - bullet list syntax correctly inside message bubbles.
• Message timestamps: Each message bubble shows a HH:MM timestamp below it (toggle via showTimestamps parameter).
• Copy to clipboard: Long-press any message bubble to copy its text to the system clipboard.
• Multi-line input: The text input field now expands up to 4 lines for longer messages before scrolling.
• Auto-focus input: The text field is focused automatically when the chat panel opens.
• bubbleStyle parameter: New AskAiBubbleStyle model lets you customize the floating button's background, icon, label color, border, and border radius.
• initiallyOpen parameter: Optionally start the widget with the panel already open.
• showTimestamps parameter: Toggle message timestamp visibility.
• maxTokens parameter: Configure the maximum response token budget for any provider (Gemini, OpenAI, Claude).
• Status indicator: Header subtitle dynamically shows "Thinking…" while a response is in-progress and reverts to "Active" when idle.
• Disabled reset during loading: The reset button is disabled while a response is streaming to prevent state corruption.
• Tooltip on bubble button: Added a tooltip to the floating bubble button for accessibility.

1.2.0

• Resolved include warning for flutter_lints in the example package analysis configuration.
• Added bubble minimization and cancel handle features to the README documentation.

1.1.0

• Added capability to minimize the floating bubble button.
• Added close/cancel button to trigger minimization.
• Added widget and unit tests verifying minimization states.

1.0.0

• Complete documentation coverage of public API.
• Restructured library entrypoint export.
• Fixed analyzer lints and unnecessary library name warnings.
• Added detailed provider support and widget layout enhancements.

0.0.1

• Initial release of ask_bubble_ai.
• Added the floating AI chat bubble widget AskBubbleAi.
• Integrated Google Gemini AI capabilities via the google_generative_ai SDK.
• Supported contextual grounding with custom knowledge base/FAQs document (brainDocument).
• Added session context awareness via user metadata mapping (userDetails).
• Implemented structured [NAVIGATE:route|text] parsing for in-app page redirection buttons.
• Added welcome screen with customizable quick question suggestion chips.
• Built beautiful dark and light mode UI matching premium design standards.