Chat Channel: Difference between revisions
| Line 19: | Line 19: | ||
Agent attributes can be used to guide the behavior of the underlying [[AI Agent]]'s [[Generative AI (GenAI)]] model. | Agent attributes can be used to guide the behavior of the underlying [[AI Agent]]'s [[Generative AI (GenAI)]] model. | ||
{{:Agent Attributes}} | |||
{{: | |||
}} | |||
===Privacy=== | ===Privacy=== | ||
Revision as of 12:54, 7 October 2026
Chat Channels define the chat interaction with AI Agents. Each channel can have multiple Chat Messages.
AI Agents
- Main article: AI Agent
The Tygron Platform has several default Chat Agents available. It is also possible to add additional Agents with custom instructions.
Instructions
An instruction is a set of prompts, settings, or rules that guides an AI Agent’s behavior, determines what data it can access, and explains how to respond to inputs.
Tools
- Main article: Agent Tool
Tools can help the Agent achieve it's goal.
Skills
- Main article: Skill
It is also possible to help the AI with Skills. These are "job descriptions" of how to do specific tasks and can increase the scope of of the agent's capabilities.
Attributes
Agent attributes can be used to guide the behavior of the underlying AI Agent's Generative AI (GenAI) model.
Agent attributes can be used to guide the behavior of the underlying AI Agent's Generative AI (GenAI) model.
| Attribute | Unit | Description | Default value |
|---|---|---|---|
| TEMPERATURE | number | Temperature controls the randomness of text that is generated by Agent's GenAI model during inference. A value >1 is more creative, <1 is more strict behavior. | 1 |
| THINKING_BUDGET_TOKENS | number | The maximum number of tokens the Agent's LLM can think before stopping. | 1000 |
| TOP_P | number | Top-P sampling selects a group of Agent GenAI model tokens whose cumulative probability is at least P. | 0.95 |
| TOP_K | number | Top-K sampling limits the Agent's LLM to choosing from the K most likely tokens at each step. | 64 |
| MAX_TOKENS | number | The maximum number of tokens the Agent's LLM can generate before stopping. | 5000 |
| TIMEOUT_SEC | sec | The maximum time allowed for the AI Agent before it is stopped. | 600 |
| THINK_MODE | boolean | With think-mode enabled the Agent's LLM returns both the reasoning and final answer. | 1 |
| PROJECT_INFO | boolean | Whether to include some basic info about the project that is accessible by the AI Agent. | 1 |
| CACHEABLE | boolean | Debug value, when deactivated the Agent Tools can no longer cache wiki pages. | 1 |
Privacy
To protect your data in the Tygron Platform, several measures have been takes. These are described in our AI Rules.
How-to's
- How to add a Chat Channel
- How to duplicate a Chat Channel
- How to set the instructions of a Chat Channel
- How to change the LLM of a Chat Channel
- How to configure a Chat Channel
- How to allow a Chat Channel to edit a project