An AI's ability to hold a coherent conversation and process info directly depends on its "context window." In Tess AI or any other platform, understanding how this window works is key to getting better results—whether you're chatting, analyzing documents, or creating agents. This article digs into what context windows are, what tokens mean, and how you can manage these elements to get the best out of the platform.
Every interaction with an AI model happens inside a "context window." Think of it as the AI's short-term memory during a specific conversation or task. This memory is measured in tokens. Tokens aren't just words—they're bits of text that AI models use to process and understand language. A word can be one token, part of a token, or even several tokens, depending on how complex it is and the language.
Each AI model available in Tess AI, like different versions of GPT or Gemini, has a maximum token capacity for its context window. For example, the GPT-4.1 Nano model can handle up to 1 million tokens, while GPT-4o Mini supports 128,000 tokens, and some Gemini models can go up to 2 million tokens.
It's important to know this token count includes both your inputs (questions, instructions, files you upload) and the outputs the AI gives back (answers, created texts).
When the number of tokens in a conversation starts getting close to or goes over the selected model's context window limit, the AI might start showing some odd behavior. This can include losing info from earlier in the chat, less coherent answers, or what's commonly called "hallucination" (making up incorrect info).
To help users out, Tess AI has added a warning that pops up when your chat is getting too long and is near the context window limit, suggesting you start a new conversation.
Step by Step: Checking a Model's Context Window
In AI Copilot, in the chat area, click the model selector.
Next to the name of your chosen model, you'll see a small info icon (an "i" inside a circle).
Click this icon to see the model details, including its "Context Window" in tokens.
Tip: If you expect a long interaction or need to provide a lot of information, go for models with bigger context windows right from the start.
One cool feature of Tess AI is the ability to analyze files like PDFs, spreadsheets (Sheets, Excel), and documents (Docs). How you upload these files into the Knowledge Base directly affects how the context window gets used.
When you add a file to the Knowledge Base, you'll see two main options in "Context Mode":
Deep Learning Mode:
Ideal for: Smaller files (e.g.: short PDFs).
How it works: The AI reads the whole file at once, bringing all its content into the context.
Credit usage: Happens just once, when reading the whole file (about 20 credits)
Context window impact: The file's entire content will take up part of the model's context window.
RAG Mode (RAG - Retrieval Augmented Generation):
Ideal for: Large files (e.g.: long PDFs, big spreadsheets).
How it works: The AI breaks the file into multiple "chunks". When you ask something, the AI searches and analyzes just the relevant pieces to answer, instead of reading the whole file every time.
Credit usage: Happens with every question that needs to search and analyze parts of the file. (usually really low)
Context window impact: Way more efficient for big files, since only the relevant pieces are loaded into the context at the time of the question, leaving most of the window available for the interaction.
Best Practice: For short documents and quick analysis, Deep Learning can be enough. For large documents or when you need to reference specific parts of a long material repeatedly, RAG Mode is a more strategic choice, as it optimizes the use of the context window and, potentially, long-term credit consumption.
If you often do repetitive tasks that need a specific set of instructions or knowledge, or if your conversations usually get pretty long and hit context limits, creating an AI Agent in AI Studio is an awesome solution.
When you create an agent, you give it detailed instructions (prompt) and can even add specific knowledge bases just once. This initial "training" gets saved in the agent.
How does this help with the context window? Every new chat you start with your custom agent kicks off with a "clean" context window for the actual interaction, but the agent already has all the knowledge and guidelines you set. So, you don’t have to repeat long instructions each time you start a chat, which saves tokens and keeps the conversations shorter and more focused.
To learn how to create an agent, check out our other article!
If, even with the above strategies, you hit the context limit in a chat and the AI suggests starting a new conversation, the history from the current chat isn't automatically carried over.
Workaround:
Copy the most important parts from your previous chat.
Start a new chat, preferably with an AI model that has a bigger context window.
Paste the copied history into the new chat to give the AI the context it needs.
Looking to the Future: Tess AI has developed an amazing feature called "Tess Memories," which lets you save specific information and contexts to reuse in different chats — making it super easy to keep the conversation going and personalize your interactions without copying and pasting manually. Totally worth a try!
