AI Chat

ArkAssist Chat is a conversational AI assistant purpose-built for research funding. It supports streaming responses with rich markdown rendering, lets you link favorited opportunities for context-aware analysis, and maintains a full searchable chat history. Whether you need quick answers about eligibility criteria or in-depth guidance on structuring a proposal section, ArkAssist Chat provides informed, contextual responses.

Chat Hub

The Chat Hub is your starting point for all conversations. It displays your previous chat history in a paginated list (6 chats per page) with the following controls:

• Search -- Filter your chat history by title to quickly locate past conversations.
• Sort -- Reorder chats by newest first, oldest first, title A-Z, or title Z-A.
• Delete -- Remove individual chats through a confirmation modal. Deleted chats cannot be recovered.
• New Chat -- Start a fresh conversation from the hub at any time.

Each chat entry shows its title and date, making it easy to find and resume previous research threads.

Starting a Conversation

Click New Chat from the Chat Hub or the ArkAssist overview to open a live chat session. Type your question in the input field and press Enter or click the send button. The assistant understands queries about:

• Funding programmes -- Specific calls, deadlines, budget ranges, and eligibility criteria across programmes like Horizon Europe, ERC, MSCA, and national schemes.
• Regulations and compliance -- Grant agreement terms, reporting requirements, financial rules, and programme-specific constraints.
• Research methodology -- Research design, impact assessment frameworks, dissemination strategies, and evaluation criteria.
• Proposal strategy -- Consortium building, work package structuring, and how to strengthen specific sections of a proposal.

Opportunity-Aware Context

One of ArkAssist Chat's most powerful features is the ability to link favorited opportunities from ArkSearch directly into your conversation. The Opportunity Selection Sidebar displays up to 8 of your favorited opportunities. When you select one:

• The assistant loads the full opportunity details and associated documentation into the chat context.
• A documentation status indicator shows whether the opportunity data has been successfully loaded.
• All subsequent questions in that conversation are answered with awareness of the specific call's requirements, evaluation criteria, budgets, and deadlines.

This means you can ask questions like "What are the key evaluation criteria for this call?" or "How should I structure the impact section?" and receive answers grounded in the actual call documentation rather than generic advice.

Live Chat Interface

The chat interface supports real-time streaming responses with rich content rendering:

• Markdown -- Headings, bold, italic, links, and nested lists render inline as the response streams.
• Syntax-highlighted code blocks -- Code snippets include language-specific highlighting and a copy-to-clipboard button.
• Mermaid diagrams -- The assistant can generate flowcharts, sequence diagrams, and other Mermaid visualisations. These render as interactive diagrams with options to view fullscreen or inspect the underlying Mermaid code.
• Tables -- Structured data renders as formatted tables within the chat.
• Images -- When relevant, the assistant can include images in responses.

Model Selection

ArkAssist Chat includes an AI model selector dropdown that lets you choose which model powers your conversation. The available models and providers (including Gemini) depend on your account tier. Higher-tier plans unlock access to more capable models with larger context windows and improved reasoning.

Select your preferred model before starting a conversation or switch models between messages within the same chat session.

Chat Controls

Each conversation provides the following controls:

• Send -- Submit your message to the assistant.
• Refresh / Regenerate -- Re-generate the assistant's last response if you want a different answer or the response was interrupted.
• Download -- Export the entire conversation as a PDF or Markdown file for offline reference or sharing.
• New Chat -- Start a new conversation without leaving the chat interface. Your current conversation is saved automatically.
• Delete Chat -- Remove the current conversation entirely.

Context Retention

Each conversation maintains full context of all previous messages within that session. You can ask follow-up questions, refine your queries, and build on earlier answers without repeating yourself.

Key points:

• Within a conversation -- The assistant remembers everything discussed. Follow-up questions like "What about the budget for that programme?" work naturally without restating which programme you mean.
• Across conversations -- Each new chat starts fresh. To continue a previous line of inquiry, reopen the existing conversation from the Chat Hub rather than starting a new one.
• Opportunity context persists -- If you linked an opportunity at the start of a conversation, that context remains available throughout the entire session.

Tips for Effective Use

• Link an opportunity first before asking detailed questions about a specific call. The context-aware responses are significantly more precise than generic answers.
• Be specific about the programme, section, or requirement you are asking about. The assistant provides more targeted advice when the scope is clear.
• Use the assistant for research before writing. Asking about evaluation criteria, common reviewer feedback, and successful proposal patterns helps you plan your approach before committing time to drafting.
• Export important conversations as PDF or Markdown so you can reference the guidance while working in your proposal editor or offline.
• Cross-reference with official sources. While the assistant draws on comprehensive funding data, always verify critical details like deadlines and budget caps against the official call documentation.