1. Sign up with GitHub OAuth or an email + password. Students: use the same GitHub account that has your GitHub Education membership for automatic verification.
2. Upload a session from /app/upload. Drop a BLF or MDF4 file. Telemera parses it in the background — large files may take a minute.
3. Open the viewer from /app/sessions. Pick signals from the tree, watch waveforms, scroll the event list.
4. Ask the AI with the Ask AI button. Questions get answered using anonymized statistics — never raw signal data.
5. Generate a report with the Generate report button. Get a private share URL (7-day expiry) that you can send to anyone without exposing the raw file.

v0.1 natively reads two formats:

• BLF — Binary Logging Format (CAN bus). Single recording with optional DBC for signal extraction.
• MDF4 — Measurement Data Format v4 (ASAM standard). Used widely by OEM and Tier 1 powertrain programs.

Additional bus log, measurement, and motorsport data formats are on the Phase 2 roadmap. Tell us which format you need via the contact form and we'll prioritize the most-requested ones.

The viewer has three resizable areas: a signal tree on the left, waveform + event list in the middle, and numeric / lap-times panels on the right. Drag the separators to taste — your layout persists for the session.

Click a signal in the tree to toggle it on the waveform. The cursor in the waveform is shared with the numeric panel — hover anywhere to read instantaneous values across all visible signals.

Click Ask AI in the viewer to open the chat sidebar. Type a question in plain English. Examples:

• “What anomalies are present in this session?”
• “Which signal had the largest excursion?”
• “Why might torque drop at 12.3 seconds?”

Telemera builds a privacy-safe payload from your session statistics and detected anomaly summaries, then streams the Claude response back. The raw time-series data never leaves the Telemera server.

From the viewer, click Generate report to render a server-side HTML snapshot of the session: stats, anomaly table, signal table. You get a private share URL with a 7-day cryptographic expiry built into the URL itself.

Sharing a report does not share the raw file — only the rendered HTML. Recipients see anomaly summaries and statistics, not the underlying time-series.

Telemera operates on the strict rule that raw telemetry data is never sent to any AI provider.

What we store on Telemera:

• The uploaded file (encrypted at rest in Cloudflare R2; you can delete it any time)
• An extracted signal index (name, ECU, unit, sample count, basic stats)
• Anomalies detected by our local statistical detectors
• Any reports you generate (rendered HTML)

What we send to the AI provider per chat turn:

• Session metadata (name, duration, signal count)
• Per-signal stats: min / mean / max / stddev
• Top anomalies (ranked by severity + score)
• Your question

Identifiers like VIN, ECU serial, and project codes are filtered out of the prompt before it leaves our servers. You can download a JSON archive of your data, or delete your account entirely, from Settings → Privacy.

Hit ? anywhere inside the app to see the live shortcut list.

• g d — Dashboard
• g s — Sessions
• g r — Reports
• g u — Upload
• g t — Settings