Tesseract

MCP Tools Reference

Every MCP tool Tesseract exposes to your AI, with example queries.

Tesseract exposes 9 tools through MCP. Your AI calls these automatically — you just ask questions in natural language.

Activity tools

get_recent_activity

Returns the most recent captured screen activity.

Example: "What was I just looking at?"

get_activity_range

Returns activity for a specific date/time range. Accepts ISO 8601 dates or local datetime strings.

Example: "Show my activity from Monday to Wednesday"

get_activity_sessions

Returns daily activity as structured work sessions with apps, projects, and content snippets. Detects session boundaries using time gaps, app changes, and semantic shifts. Designed for LLM consumption — returns the full day without truncation.

Example: "Break down my workday into sessions"

get_app_usage

Shows which apps you used and for how long.

Example: "Which apps did I use most today?"

Search tools

search_activity

Full-text keyword search over your captured text (window titles, OCR content, URLs).

Example: "Find mentions of API design"

Vector similarity search — finds activity related to a concept, even without exact keyword matches.

Example: "What was I working on related to authentication?"

Summary tools

summarize_day

Generates an AI-powered summary of daily activity, grouped by project with time spent. Uses the configured LLM provider to produce a time-tracking-friendly summary from detected work sessions.

This tool sends captured text to your configured LLM provider. If using Ollama or Apple Intelligence, everything stays local. If using the Claude API, text is sent to Anthropic for processing.

This tool only appears when an LLM provider is available — either a Claude API key is configured, Ollama is set up, or you're running macOS 26+ (Apple Intelligence). See Configuration for setup instructions.

Example: "Summarize my day for a timesheet entry"

Stats tools

get_stats

Returns database statistics — total captures, date range, storage size.

Example: "How many captures are in the database?"

get_llm_status

Shows which LLM providers are configured and available for summaries.

Example: "What LLM providers are set up?"

On this page