Hello, Memfit!
This chapter is a basic orientation for Memfit. It does not explain every feature in depth; it gives you the first map: what the main interface contains, how to start your first task, and where to go next for knowledge bases, Skills, and advanced usage.
【Image: Memfit AI Agent homepage overview, labeling chat area, resources, model/usage entry, and settings entry】
What Memfit Is
Memfit is an AI Agent workspace that can execute tasks. It can plan, call tools, request permissions, use context, and show the execution process instead of only returning a final answer.
| Capability | What You See | Learn More |
|---|---|---|
| Agent execution | Chat, plans, tool calls, Review confirmations, results | Agent Tutorial |
| Knowledge base | Knowledge lists, document import, retrieval, mounting | Knowledge Base Tutorial |
| Skills | Skill list, selection, parameter confirmation, import/export | Loading and Using SKILLS |
| Advanced usage | Usage statistics, context pressure, model performance, custom tools | Advanced Usage |
First Look at the Main Interface
| Area | Purpose |
|---|---|
| Chat area | Enter tasks, view replies, and follow execution. |
| Resource area | Manage knowledge bases, Skills, tools, and other resources. |
| Model and usage entry | View current models, context tokens, input/output usage, pressure, and speed. |
| Review panel | Confirm plans, tool calls, or Skill execution. |
| Settings entry | Adjust basic settings, permissions, AI model configuration, and calling mode. |
【Image: Main interface area labels】
Send Your First Task
Try a small task:
Hello, introduce what Memfit can help me do in three sentences.
Watch whether Memfit answers directly, creates a plan, asks for Review, or references tools, knowledge bases, or Skills.
【Image: First message and reply】
Understand the Execution Process
| Event | Meaning | What You Do |
|---|---|---|
| Plan | AI breaks down a complex task. | Check whether the direction is correct. |
| Tool call | AI is preparing to use a tool or system capability. | Review the target, path, and parameters. |
| Observation | A tool returns data and AI continues analysis. | Wait or add information if requested. |
| Final answer | AI summarizes findings or outputs deliverables. | Validate and continue the task if needed. |
【Image: A task showing plan, tool call, observation, and final answer】
Which Tutorial to Read Next
Read Agent Tutorial if you want to execute tasks, attach files, approve Reviews, and understand results.
Read Knowledge Base Tutorial if you want AI to answer based on your documents, standards, vulnerability data, or internal materials.
Read Loading and Using SKILLS if you want scenario-based workflows such as code audit, vulnerability analysis, or report generation.
Read Advanced Usage if you want to inspect usage, tune cost and performance, or customize tools and SKILLs:
- Usage and Model Performance: inspect token usage, cache hits, context pressure, speed, and jitter.
- Manual Intervention and Context Editing: redirect execution, add constraints, and distinguish global context from local context.
- Custom Tools and Custom SKILLs: package team scripts, fixed workflows, and reusable capabilities.
Recommended Learning Path
- Finish Installation.
- Finish Basic Settings and Custom AI Configuration.
- Read this chapter to understand the interface.
- Follow Agent Tutorial to execute a real task.
- Continue with knowledge bases, Skills, and advanced usage as needed.