What is context?
Context is your personal setting for the AI assistant. Without context, the AI doesn’t know your role, experience, technologies, and response format preferences. With context, it responds exactly as you need.
You can create multiple contexts for different types of interviews:
- HR interview - focusing on soft skills and behavioral questions
- Technical interview - emphasizing technologies and code examples
- and others
Switch between them before the interview with one click.
Context parameters
Context name
A name for quick identification of the preset in the list. For example: “Python Backend”, “React Frontend”, “DevOps Engineer”.
Custom context
A toggle for advanced users. If enabled, all other fields are ignored, and only your text from the block below is used.
Use this if you want full control over instructions for the AI.
Role
The position you’re applying for. Examples:
- Senior Python Developer
- Frontend React Developer
- DevOps Engineer
- Data Scientist
The AI adapts responses to the specified role: for a Frontend developer it will talk about React and TypeScript, for DevOps - about Kubernetes and CI/CD.
Response language
The language in which the assistant will respond. Available:
- Auto-detect - AI will determine the language from context
- Russian, English, and 50+ other languages
Useful if the interview is in English, but you want to see hints in Russian (or vice versa).
Resume (PDF)
Upload a PDF resume - the text will be extracted and added to the context. The AI will use your real experience when answering.
How it works:
When the interviewer asks “Tell me about project X”, the AI already knows what you did there and responds based on your real experience.
Personal data (name, contacts) is automatically filtered out when processing the PDF.
Technologies
A list of technologies you work with. Enable the toggle and add technologies through the input field.
How it affects responses:
If you specify Python, FastAPI, PostgreSQL - the AI won’t mention Java or MongoDB in responses. All code examples will be in the specified technologies.
Additional data
Any additional information for the AI:
- Specific facts about your experience
- Special instructions for responses
- Context about the company you’re interviewing with
Example:
- Always mention metrics and specific numbers
- Use examples from the fintech domain
- Don't mention the previous employer by name
Response length
Controls the volume of generated responses:
| Value | Description |
|---|
| Short | 100-600 characters. 1-2 sentences or a brief list. For simple questions. |
| Medium | 700-1500 characters. Balance between brevity and details. |
| Extended | 1500+ characters. Maximum details, examples, and explanations. |
Defines the structure of responses to your direct questions in chat:
| Format | Description | When to use |
|---|
| Terms | Clear definitions and explanations | Technical questions |
| Terms and examples | Definitions + code examples | Questions with code |
| STAR | Situation-Task-Action-Result | Behavioral questions |
| Problem-Action-Result | Problem-Action-Result | Case interview |
| Custom | Your own template | Specific requirements |
When selecting “Custom”, a field will appear to describe your format.
A similar setting, but applies to responses to messages transcribed from voice (system audio and microphone transcriptions).
You can configure differently: for example, STAR for direct questions, and brief terms for transcriptions.
Transcription language
Language for speech recognition. Affects transcription accuracy.
Choose the language in which the interview is conducted. If the interview is in English - set English.
Keywords for transcription
Specific terms, technology names, abbreviations that should be recognized more accurately.
Example:
React, TypeScript, Kubernetes, PostgreSQL, CI/CD, gRPC, Kafka
These words will be prioritized during speech recognition, reducing errors in technical terms.
Setup recommendations
-
Configure context before the interview - spend 5 minutes, get relevant responses.
-
Upload an up-to-date resume - the more accurate the data, the better the quality of responses.
-
Specify real technologies - don’t add what you don’t know.
-
Create multiple contexts - different profiles for different types of interviews.
-
Use additional data - this is fine-tuning for natural responses.