Chat Info

Pentos has two info panels: Chat Info for conversation-level details and Message Info for per-response metadata. Both slide in from the right side of the screen.

Chat Info

Open the chat options menu (the + button at the top left of the chat) and select Info to open the Chat Info panel.

Details

  • Chat Name: The current title. Click the edit icon next to it to rename the chat from this panel.
  • Created: When the conversation was started.
  • Last Updated: When the most recent message was sent.

Usage

  • Total Messages: The number of messages in the conversation, including both your messages and AI responses. Deleted messages aren’t counted, and neither are activity notes such as someone joining or the chat being renamed.
  • Input Tokens: Total tokens sent to the AI across all responses, including conversation history and context.
  • Output Tokens: Total tokens generated by the AI in its responses.
  • Reasoning Tokens: Total tokens used for internal reasoning by models that support it.
  • Cache Write Tokens and Cache Read Tokens: Tokens written to or served from the provider’s prompt cache, which lowers cost and energy on later messages.
  • Estimated Energy: The estimated total energy consumed by the conversation. See Energy Usage for how estimates are calculated.
  • Estimated CO₂: The approximate carbon footprint of that energy, recorded for each response at the grid average where it was generated.
  • Estimated Water: The approximate water behind that energy: data-center cooling plus the water used to generate the electricity. See Energy Usage.

Members

The people the chat is shared with and their roles. If nobody else has access, this section says “Not shared with anyone.” When there’s more than one, click their photos to show the list. Click someone to open their profile in the User Profile panel, or use the buttons beside them: Add friend, when you aren’t friends and haven’t sent them a request, and Message, which opens your direct message with them in Collab. Your own row isn’t clickable and has no buttons. See Sharing & Permissions.

Associations

Files, chats, tasks, and other items linked to this chat.

Message Info

Open the message menu on an AI response (the + button beneath the AI provider’s logo, or right-click the response) and select Info.

Model

  • Provider: Which AI provider handled the response.
  • Model: The specific model used (for example, GPT 6 Luna or Claude Sonnet 5.5).
  • Smart Routed: Shown when the Pentos Smart Router picked the model. Click it to see the Smart Router’s routing details for that response.
  • Personality: The AI Personality that was active for this response.
  • Agent Tasks: Shown when the AI posted a task list for the response. Click it to view the plan.

Preferences

Shows which chat controls were active for this response:

  • Personalization: Whether your profile information was sent with the message.
  • Location: Whether your approximate location was shared with the AI.
  • Web Search: Whether the AI was allowed to search the web for live information.
  • Use My Key: Whether the response ran on your own API key.
  • Context Mode: Which context mode produced the response: Lean (last 10 messages), Standard, or Max.
  • Effort: The effort level used for the response.
  • Connectors: Whether the AI had access to your connected apps.

Usage

  • Points: The points this response cost. Click the number to see what they were for: the model’s points, each chat setting that added to them, and any image the AI made or edited for the response. Images made with your own API key cost no points, so they aren’t listed.
  • Input Tokens: Tokens from your prompt and conversation history sent to the AI.
  • Output Tokens: Tokens in the AI’s response.
  • Reasoning Tokens: Tokens used for internal reasoning by models that support it.
  • Cache Write Tokens and Cache Read Tokens: Shown when the response wrote to or read from the provider’s prompt cache.
  • Connectors: Shown when the AI used connected apps or tools. Click View to see which apps and tools ran.
  • Estimated Energy: The estimated energy consumed by this response. See Energy Usage for details.
  • Estimated CO₂: The approximate carbon footprint of that energy, at the grid average where the response was generated.
  • Estimated Water: The approximate water behind that energy: data-center cooling plus the water used to generate the electricity. See Energy Usage.
  • Estimated Cost: Shown only when the response ran on your own API key: the estimated cost in USD at the provider’s published prices.

Citations

If the AI used web sources, links to those sources appear here. If no web sources were used, this section shows “No citations.”

Content

Any files attached to this message are listed with their names and sizes. Click a file to open a preview. If the message had no attachments, this section shows “No attachments.”

References

The items the AI read while writing this response, such as files, chats, channels, conversations, and Memories. Click a file, folder, chat, channel, conversation, or memory to open it. If the AI did not read anything, this section shows “No references.”

Chat ActionsBring Your Own Key