The recap Agent Week · July 21–23

AI agents built for
robotics data.

Stop replaying logs one at a time. Roboto Agents flag events, triage failures, and automate analysis across your whole fleet.

Meet Roboto Agents
Tue Jul 21
Agents. Live
Wed Jul 22
Triage. Live
Thu Jul 23
Close the loop. Live
Day 01 / Agents

Agents that understand robotics data.

Turn every robot run into answers. Roboto Agents automate the triage, investigation and data curation tasks that used to cost you hours.

Read the post
Tools signals · video · code Skills runbooks & know-how Context robot · mission · code Data model every format, queryable
flight_0349.ulg px4-roboto#101 Open [Triage] Severe IMU clipping → EKF fault Roboto::high_vibration Roboto::ekf_estimator_failure +1 Triage Agent
Day 02 / Triage

Triage every log automatically.

Roboto Agents analyze every log the moment it arrives and open a labeled, evidence-backed issue in Linear, GitLab, or Jira.

Read the post
Day 02 / Triage

Ask one run, or ten thousand, anything.

AI Chat analyzes your multimodal data, plots key signals, inspects video and highlights the moments that matter.

Read the post
vehicle_gps_position 3d_scene altitude_msl_m console_logs + - 141 80 27 Time message 01:31:26.8 [mc_pos_ctl] arm 01:31:26.9 Failsafe activated 01:31:35.1 [timesync] offset 01:31:43.8 [ekf2] primary reset 01:32:40.8 [gps] ubx nav 01:32:41.8 [gps] ubx sat 01:32:44.5 [mc_pos_ctl] land What went wrong on this flight? Used get_events Used plot_altitude Used search_logs Several critical anomalies occurred during this flight: Estimator Filter Fault Primary EKF filter fault Bad vertical accel ~65s Blind Land Failsafe Failsafe: blind land mode Emergency landing
Day 03 / Close the loop

Trace a robot failure into the code that caused it.

Roboto Agents test hypotheses against your data, follow the evidence into your repo through MCP, and come back with a cause and a fix to review.

Read the post
MAGNETOMETER · Z flight_0349 14:32 · anomaly ekf2_params.cpp +1−1 est-config · main 41 42 − + 44 45
grasp_17 grasp_06 grasp_14 grasp_02 grasp_10 grasp_08 grasp_16 grasp_04 grasp_13 grasp_24 grasp_42 grasp_26 grasp_15 grasp_31 grasp_07 grasp_19 grasp_03 grasp_11 grasp_22 grasp_09 grasp_36 grasp_18 grasp_28 grasp_44 grasp_12 grasp_05 grasp_33 grasp_21 grasp_01 trainsuccessful_grasps512 clips valfailed_grasps148 clips traincluttered_scenes207 clips trainnovel_objects96 clips Export to PyTorch Hugging Face S3
Day 03 / Close the loop

Describe a moment, and Agents find every example.

Roboto Agents search your time-series, video, and logs for the moments you describe, then gather every match into a Collection to review, share, or train on.

Read the post