Use Cases

Common patterns for building with m8tes. Each section shows what to do and links to a working example.

Internal team automation

Automate recurring ops work: reports, triage, monitoring, and scheduled runs.

Weekly reports with file generation

Generate markdown summaries and CSV exports on a schedule, then download the results.

See examples/file-report.py for a full pipeline: streaming progress, file tracking, and download.

Support triage with streaming

Stream a run in real time and filter events to show only what matters.

See examples/support-triage.py for a complete real-time event filter example.

Revenue monitoring and alerts

Run analyses on a schedule and post results to Slack automatically.

Embed in your product

Run agents on behalf of your users with per-user memory, OAuth, and data isolation.

Multi-tenant agents with per-user memory

Each user gets isolated context: memory, task history, and tool connections are separate.

See Users for full isolation docs.

OAuth connection flow per user

Connect integrations (Gmail, Slack, etc.) per end-user and guard runs until connected.

See examples/embed-oauth.py for a complete per-user OAuth flow with connection guards.

Streaming chat interface in your app

Build a live chat UI by streaming text chunks to the frontend.

Plan and approve

Let the agent propose a plan before acting, or require approval before each tool use.

Interactive terminal approval

The agent pauses before each tool call, you approve or deny interactively.

See examples/plan-mode.py for a complete interactive and automated approval example.

Automated approval in production

Auto-approve safe tools, deny dangerous ones, no human needed.

Webhook-based approval for async apps

Receive permission requests as webhook events, respond via the API.

See Human-in-the-Loop for the full webhook-based flow.

Generate reports and files

Produce structured output files (CSV, JSON, Markdown) and download them after the run.

Stream progress and download

Watch file writes in real time, then download when done.

See examples/file-report.py for the complete version.

Scheduled weekly pipeline

Combine file generation with scheduling for fully automated reports.

Multi-format output

Ask the agent to write multiple file types in a single run.