Skip to main content
Conversations in ParleHub are not private chat windows — they are shared, persistent threads that every project member can read, contribute to, and search. When a colleague sends a message, you see it in real time. When the AI responds, everyone watching sees the reply arrive token by token. This shared visibility means your team builds on each other’s work instead of duplicating it, and the full reasoning behind every decision stays in the thread for the life of the engagement.

​
Starting a conversation

1

Open the project

Select the project from your sidebar. You land on the project’s Conversations tab, which lists all existing threads.
2

Create a new thread

Click New Conversation in the top-right corner. Give the conversation a descriptive title so teammates can find it later — for example, “Contract risk review — MSA v3”.
3

Select a model

Choose an AI model from the model picker. The dropdown shows only the models your project admin has approved for this engagement.
4

Send your first message

Type your message and press Enter (or click Send). The thread is now live and visible to all project members.

​
Real-time collaboration

ParleHub shows you exactly what your teammates are doing in a conversation at any moment, so you never send a duplicate message or interrupt a reply that is already in flight.
  • Viewing — A subtle avatar badge appears next to the conversation title when a colleague has the thread open.
  • Typing — A live indicator shows when someone is composing a new message, so you know a follow-up is coming before you ask the same question.
  • Waiting on a reply — When a message has been sent and the AI is still responding, teammates see a streaming indicator rather than a blank thread, giving everyone a clear signal that work is in progress.
All updates arrive without a page refresh. If you have the conversation open in one tab and a colleague sends a message in another browser, your view updates instantly.

​
Branching

Branching lets you explore an alternative direction in a conversation without destroying what you already have. To create a branch, hover over any message you sent, click Edit, revise the text, and press Resend. ParleHub keeps the original message and its downstream replies intact as a separate branch, and the new version opens as a parallel path from that same point. This is fundamentally different from deleting a message and retyping: deleting removes the original exchange permanently, which can wipe out useful context. Branching preserves both paths, so you can compare outputs, revisit the original answer, or share both branches with a colleague for review.
Branches are visible to all project members. Use the branch selector at the fork point to switch between the original thread and any alternatives.

​
Promoting a personal chat

If you started exploring an idea in a personal (private) chat and it has turned into something worth sharing with the engagement team, you can promote that conversation into a project thread without copy-pasting. Open the personal conversation, click the ⋯ menu in the top-right corner, and select Promote to Project. Choose the destination project and confirm. The full conversation history — including the AI’s responses — moves into the project and becomes visible to all members immediately.
Promotion is a one-way operation. Once a personal chat is promoted to a project, it is no longer private. Review the thread before promoting to make sure it contains only content appropriate for the full project team.
Use the Search bar at the top of the Conversations tab to find anything that has been discussed in the project. ParleHub supports two search modes:
  • Keyword search matches exact words and phrases across all conversation messages, including AI responses.
  • Semantic search understands meaning and intent, so searching for “liability cap” can surface messages that discuss “maximum damages” or “exposure ceiling” even if those exact words never appear.
Both modes search the full history of every conversation in the project, not just the one you are currently viewing.
Before starting a new conversation, run a quick semantic search for your question. If a colleague already worked through the same issue last week, you get the answer in seconds rather than spending tokens — and budget — re-deriving it from scratch.