Your first session
Launch the workspace, run an agent, read the numbers. What you'll see and how to interpret it.
Open the workspace
senkani
Or open SenkaniApp from /Applications if you dragged it there. A single empty canvas appears.
Open three panes
⌘K → type "Terminal" → Enter. ⌘K → "Analytics". ⌘K → "Agent Timeline". Three panes side by side.
Start the agent
In the Terminal pane:
claude
Claude Code starts; MCP handshake happens automatically (you can confirm via the "MCP connected" indicator in the status bar).
Ask for something real
Ask the agent to do something that involves reading your project:
> Find the function that handles order pagination.
Watch the Agent Timeline pane fill with tool calls (senkani_search → senkani_fetch → possibly senkani_read). Each line shows the tool, target, and compression ratio.
Read the Analytics pane
- Tokens saved — bytes that would have been sent to the LLM but weren't.
- Cost saved — tokens × model rate (from the pane's model preset).
- Compliance rate — fraction of tool calls that went through Senkani vs the agent's built-ins. High is good.
- Per-feature breakdown — F / C / S / I / T contributions.
What "good" looks like
- Compliance > 80% — the agent is using
senkani_*tools. - Live multiplier > 3× — you're saving meaningful tokens.
- Agent Timeline is mostly
senkani_*entries, not raw Read/Bash/Grep.
If compliance is low
The agent is bypassing senkani tools. Run senkani doctor to check registration; make sure the MCP server entry in ~/.claude/settings.json is present; make sure you launched claude inside a Senkani-managed Terminal pane (the one the app opened, not a random shell).