diff --git a/research/expansion/MUJOCO_VIZ_ANALYSIS.md b/research/expansion/MUJOCO_VIZ_ANALYSIS.md
new file mode 100644
index 0000000..1173ab0
--- /dev/null
+++ b/research/expansion/MUJOCO_VIZ_ANALYSIS.md
@@ -0,0 +1,141 @@
+# opentrons-mujoco-viz: source analysis and measured behaviour
+
+Reviewed: `PhysicalAIBenchmarks/opentrons-mujoco-viz` at commit `460169f` (about 1,275 lines: `scene.py` 352,
+`event_driver.py` 249, `generate_labware.py` 197, `modules.py` 180, `assets/ot2.xml` 246), with MuJoCo 3.15.
+Every number below was measured by `tools/probe_mujoco_viz.py`; raw output is in `mujoco_viz_probe.json`.
+Evidence for PR #63, which rendered all tasks and candidates with this package.
+
+## Summary
+
+The package draws a recognisable OT-2 and keeps correct liquid and module state, and it renders a video in about 5 seconds.
+But the simulated gantry does not go where it is commanded: the pipette tip ends 17–103 mm from the centre of every slot,
+never changes height, and never visits an individual well. The MuJoCo part therefore adds motion that looks plausible but is
+not a faithful account of the protocol, and it cannot yet support collision or reach checks.
+
+## How it works
+
+| Part | What it does |
+|---|---|
+| `assets/ot2.xml` | A hand-built scene of 17 bodies and 82 box/cylinder geoms, no meshes. One fixed deck: tip rack (slot 1), plate (2), trough (3), magnetic module (4), temperature module (6), thermocycler (7/8), heater-shaker (9), trash (12). Three slide joints (x, y, z) with position actuators, plus the thermocycler lid. |
+| `modules.py` | `ProtocolState`: pure-Python bookkeeping of pipette volume, tip, per-well volumes and module states, updated by each event. The docstring in `scene.py` says so directly: "Liquid volumes are tracked analytically; MuJoCo handles gantry kinematics only." |
+| `scene.py` | `WetLabScene.apply_event` moves the gantry to the event's slot for a fixed number of physics steps, updates the state, recolours geoms, and returns frames. `render_with_hud` overlays step count, tip, pipette volume and module status. |
+| `event_driver.py` | Regex parser for `opentrons_simulate` text output and a `render_protocol` wrapper. |
+| `generate_labware.py` | Builds OBJ meshes from Opentrons labware definitions. Not used by the bundled scene (0 meshes loaded). |
+
+## Findings
+
+### 1. The gantry does not reach the commanded position
+
+Tip position after `move_to_slot(slot, work=True)` plus six further seconds of simulated time:
+
+| Slots | Row | Tip distance from slot centre |
+|---|---|---|
+| 1, 2, 3 | front | 102–103 mm |
+| 4, 5, 6 | second | 23–44 mm |
+| 7, 8, 9 | third | 18–45 mm |
+| 10, 11, 12 | back | 17–40 mm |
+
+Three separate causes:
+
+- **Y is out of range for the front half of the deck.** `_slot_ctrl_y` commands `slot_y − 0.17`, which is −0.34 m for slots 1–3
+  and −0.255 m for slots 4–6. The joint and actuator range is ±0.24 m, so the command is clamped and the tip stops at y = −0.07 m
+  for all six of those slots. The pipette never reaches the front row.
+- **X stops 17–45 mm short.** With contacts disabled the x joint reaches its target exactly (−0.152 commanded, −0.152 achieved),
+  so the shortfall comes from collisions.
+- **Motion is not finished when frames are taken.** `apply_event` allows 100 physics steps (0.2 s) per move. After that budget the
+  x joint was still at −0.013 m for a −0.152 m command (slot 1) and at −0.121 m for a 0 m command (slot 2), so rendered frames show
+  the gantry part-way to, or still leaving, the previous slot.
+
+### 2. The pipette never changes height
+
+`SAFE_Z` is −0.02 m and `WORK_Z` is −0.17 m, but the z joint sits at −0.354 m and −0.369 m respectively, and the tip is at
+z = −0.152 m over every slot. The two heights differ by 15 mm on screen instead of 150 mm. The actuator is pushing with about
+2,700 N against 16 active contacts, and turning gravity off does not free it. With contacts off as well the joint goes to −0.517 m,
+beyond its stated −0.38 m limit, so the limit is not being enforced either. The cause of the downward drift was not isolated.
+
+### 3. Collisions are on for every geom, and they are the wrong ones
+
+All 82 geoms have collision enabled and there are 12–16 active contacts at all times, including at rest. These are the model's own
+parts touching each other, not the tip touching labware. Until they are removed, contact counts say nothing about a protocol.
+
+### 4. Wells are not targeted
+
+`move_to_slot` accepts a `well_offset_x`, but `apply_event` never passes one and there is no y offset at all, so the commanded
+target for well A1 and well H12 of the same plate is the same point, the slot centre. Every operation on a plate is aimed at one spot.
+(This is read from the code; the measured joint positions for the two wells differ only because the gantry had not settled.)
+
+### 5. Only eight wells show liquid
+
+The state tracks volume for any slot and well (100 µL dispensed into slot 5 C7 is recorded), but the scene only recolours eight
+geoms, `w_A1` to `w_H1` in slot 2. Liquid anywhere else on the deck is invisible. This is why the four panels of a
+model-versus-oracle grid look the same.
+
+### 6. Liquid bookkeeping accepts impossible volumes
+
+- Aspirating 5,000 µL leaves 5,000 µL in the pipette; `PIPETTE_MAX = 1000.0` is defined but never checked.
+- Wells start at 0 and `add_to_well` clamps at 0, so aspirating from a source leaves it at 0. Source volumes are never tracked and
+  drawing from an empty well is not detected.
+- The tip-liquid indicator saturates at 200 µL regardless of pipette size.
+
+### 7. Four of the five named cameras render black
+
+Mean pixel brightness (0–255): `iso` 0.0, `side` 0.0, `front` 0.0, `deck` 0.03, `top` 87.8, and the built-in fallback free camera 31.0.
+`render()` and `apply_event()` default to `iso`, so the documented quick start produces black frames. `scripts/render_run.py`
+in this repository already works around this by forcing the fallback camera.
+
+### 8. The package's parser does not read current Opentrons output
+
+On the `a1-a12-100ul` reference protocol under Opentrons 7.5, the log has 12 aspirate and 12 dispense lines; `parse_simulate_output`
+returns 2 events (one pick, one drop) and no liquid moves. Its pattern expects `well A1 of <labware> on <slot>`; Opentrons 7.5 prints
+`Aspirating 100.0 uL from A1 of reservoir on 1 at 92.86 uL/sec`. `render_run.py` and `tools/mj_render.py` avoid it by converting
+events from `eval/runlog.py`.
+
+### 9. The deck is the same for every protocol
+
+The scene has no notion of what a protocol loaded. A protocol with tube racks in slots 1 and 2 is drawn with a tip rack and a
+96-well plate there. `generate_labware.py` could supply real geometry but is not wired in.
+
+## What works
+
+- State tracking of pipette, tip, well volumes and module states is straightforward and correct for well-formed logs.
+- Module feedback: magnet bar and LED, temperature LED colour, thermocycler lid actuation, heater-shaker latch.
+- Speed: about 5 s for a 45 s video, headless, with no browser or GPU.
+- The HUD shows step count, tip state, pipette volume and elapsed protocol time.
+
+## Compared with the PyLabRobot replay
+
+| | opentrons-mujoco-viz | PyLabRobot replay (`tools/replay_log.py`) |
+|---|---|---|
+| Deck | One fixed layout | Built from the labware the protocol loaded |
+| Position shown for an operation | 17–103 mm from the slot centre; wells not distinguished | The exact well, from the labware definition |
+| Height changes | 15 mm instead of 150 mm | Marker above the well (no z motion modelled) |
+| Liquid shown | Eight wells of one plate | Every well and tube |
+| Tips | One indicator | Every tip position |
+| Faults | None | Marked by `fault_detect.py` |
+| Robots | OT-2 | OT-2, and a Hamilton STARlet |
+| Physics | Rigid-body dynamics, currently obstructing the motion | None |
+| Render time per video | About 5 s | 1–2 min, needs a GPU browser |
+
+Side-by-side videos of all 11 tasks, both driven by the same oracle run log, are in the private Hugging Face dataset
+`EvanOLeary/text2wetlab-3d-replays` under `side_by_side/` (built by `tools/side_by_side.sh`).
+
+## What to fix, in order
+
+1. Turn collisions off for the gantry and decorative geoms (`contype="0" conaffinity="0"`), or add contact exclusions, so the
+   gantry can move. This alone should fix x tracking.
+2. Correct the y mapping or widen the y range so slots 1–6 are reachable.
+3. Find why z does not hold its commanded height, and enforce the joint limit.
+4. Step until the position error is below a tolerance instead of a fixed 100 steps.
+5. Pass per-well offsets from the labware definition into `move_to_slot`.
+6. Build the deck from the protocol's labware (wire in `generate_labware.py`) and recolour every well.
+7. Fix the named cameras and make a working one the default.
+8. Update the parser to the Opentrons 7.x log format, or drop it in favour of `eval/runlog.py`.
+9. Enforce pipette capacity and allow starting volumes.
+
+Once 1–5 hold, contacts between the tip and labware become meaningful, and a collision or reach check can be added to
+`fault_detect.py` as a new fault type. That is the capability MuJoCo offers that the PyLabRobot replay cannot.
+
+## Limits of this analysis
+
+One commit of the package was examined with one MuJoCo version. The z-axis behaviour was measured but its cause was not isolated.
+The tip position is the centre of the `pip_tip` geom; slot centres are the package's own `SLOT_XY` table.
diff --git a/research/expansion/mujoco_viz_probe.json b/research/expansion/mujoco_viz_probe.json
new file mode 100644
index 0000000..c969d5c
--- /dev/null
+++ b/research/expansion/mujoco_viz_probe.json
@@ -0,0 +1,451 @@
+{
+ "model": {
+  "bodies": 17,
+  "geoms": 82,
+  "joints": 4,
+  "actuators": 4,
+  "cameras": 5,
+  "meshes": 0,
+  "timestep_s": 0.002,
+  "geoms_with_collision_enabled": 82
+ },
+ "joint_ranges_m": {
+  "j_tc_lid": [
+   0.0,
+   0.06
+  ],
+  "j_x": [
+   -0.29,
+   0.29
+  ],
+  "j_y": [
+   -0.24,
+   0.24
+  ],
+  "j_z": [
+   -0.38,
+   0.02
+  ]
+ },
+ "camera_mean_brightness_0_255": {
+  "iso": 0.0,
+  "side": 0.0,
+  "top": 87.83,
+  "front": 0.0,
+  "deck": 0.03,
+  "free (fallback)": 30.95
+ },
+ "gantry_by_slot": {
+  "1": {
+   "commanded_joint_xyz": [
+    -0.152,
+    -0.34,
+    -0.17
+   ],
+   "joint_xyz_after_apply_event_budget": [
+    -0.0127,
+    -0.1546,
+    -0.371
+   ],
+   "joint_xyz_after_6s_more": [
+    -0.13,
+    -0.24,
+    -0.3688
+   ],
+   "slot_centre_xy": [
+    -0.152,
+    -0.17
+   ],
+   "tip_xy": [
+    -0.13,
+    -0.07
+   ],
+   "tip_error_mm": 102.4,
+   "tip_z": -0.152,
+   "contacts": 12
+  },
+  "2": {
+   "commanded_joint_xyz": [
+    0.0,
+    -0.34,
+    -0.17
+   ],
+   "joint_xyz_after_apply_event_budget": [
+    -0.1214,
+    -0.24,
+    -0.3685
+   ],
+   "joint_xyz_after_6s_more": [
+    -0.019,
+    -0.24,
+    -0.3688
+   ],
+   "slot_centre_xy": [
+    0.0,
+    -0.17
+   ],
+   "tip_xy": [
+    -0.019,
+    -0.07
+   ],
+   "tip_error_mm": 101.8,
+   "tip_z": -0.152,
+   "contacts": 12
+  },
+  "3": {
+   "commanded_joint_xyz": [
+    0.152,
+    -0.34,
+    -0.17
+   ],
+   "joint_xyz_after_apply_event_budget": [
+    -0.0014,
+    -0.24,
+    -0.3685
+   ],
+   "joint_xyz_after_6s_more": [
+    0.1279,
+    -0.24,
+    -0.3688
+   ],
+   "slot_centre_xy": [
+    0.152,
+    -0.17
+   ],
+   "tip_xy": [
+    0.128,
+    -0.07
+   ],
+   "tip_error_mm": 102.8,
+   "tip_z": -0.152,
+   "contacts": 12
+  },
+  "4": {
+   "commanded_joint_xyz": [
+    -0.152,
+    -0.255,
+    -0.17
+   ],
+   "joint_xyz_after_apply_event_budget": [
+    0.1064,
+    -0.24,
+    -0.3687
+   ],
+   "joint_xyz_after_6s_more": [
+    -0.1115,
+    -0.24,
+    -0.3688
+   ],
+   "slot_centre_xy": [
+    -0.152,
+    -0.085
+   ],
+   "tip_xy": [
+    -0.111,
+    -0.07
+   ],
+   "tip_error_mm": 43.7,
+   "tip_z": -0.152,
+   "contacts": 12
+  },
+  "5": {
+   "commanded_joint_xyz": [
+    0.0,
+    -0.255,
+    -0.17
+   ],
+   "joint_xyz_after_apply_event_budget": [
+    -0.1069,
+    -0.24,
+    -0.3685
+   ],
+   "joint_xyz_after_6s_more": [
+    -0.0168,
+    -0.24,
+    -0.3688
+   ],
+   "slot_centre_xy": [
+    0.0,
+    -0.085
+   ],
+   "tip_xy": [
+    -0.017,
+    -0.07
+   ],
+   "tip_error_mm": 22.7,
+   "tip_z": -0.152,
+   "contacts": 12
+  },
+  "6": {
+   "commanded_joint_xyz": [
+    0.152,
+    -0.255,
+    -0.17
+   ],
+   "joint_xyz_after_apply_event_budget": [
+    0.0003,
+    -0.24,
+    -0.3685
+   ],
+   "joint_xyz_after_6s_more": [
+    0.1282,
+    -0.24,
+    -0.3688
+   ],
+   "slot_centre_xy": [
+    0.152,
+    -0.085
+   ],
+   "tip_xy": [
+    0.128,
+    -0.07
+   ],
+   "tip_error_mm": 28.3,
+   "tip_z": -0.152,
+   "contacts": 12
+  },
+  "7": {
+   "commanded_joint_xyz": [
+    -0.152,
+    -0.17,
+    -0.17
+   ],
+   "joint_xyz_after_apply_event_budget": [
+    0.1072,
+    -0.2086,
+    -0.3698
+   ],
+   "joint_xyz_after_6s_more": [
+    -0.1075,
+    -0.17,
+    -0.3708
+   ],
+   "slot_centre_xy": [
+    -0.152,
+    0.0
+   ],
+   "tip_xy": [
+    -0.107,
+    -0.0
+   ],
+   "tip_error_mm": 45.0,
+   "tip_z": -0.154,
+   "contacts": 16
+  },
+  "8": {
+   "commanded_joint_xyz": [
+    0.0,
+    -0.17,
+    -0.17
+   ],
+   "joint_xyz_after_apply_event_budget": [
+    -0.104,
+    -0.17,
+    -0.3706
+   ],
+   "joint_xyz_after_6s_more": [
+    -0.018,
+    -0.17,
+    -0.3708
+   ],
+   "slot_centre_xy": [
+    0.0,
+    0.0
+   ],
+   "tip_xy": [
+    -0.018,
+    -0.0
+   ],
+   "tip_error_mm": 18.0,
+   "tip_z": -0.154,
+   "contacts": 16
+  },
+  "9": {
+   "commanded_joint_xyz": [
+    0.152,
+    -0.17,
+    -0.17
+   ],
+   "joint_xyz_after_apply_event_budget": [
+    -0.0009,
+    -0.17,
+    -0.3706
+   ],
+   "joint_xyz_after_6s_more": [
+    0.1264,
+    -0.17,
+    -0.3693
+   ],
+   "slot_centre_xy": [
+    0.152,
+    0.0
+   ],
+   "tip_xy": [
+    0.126,
+    -0.0
+   ],
+   "tip_error_mm": 26.0,
+   "tip_z": -0.152,
+   "contacts": 16
+  },
+  "10": {
+   "commanded_joint_xyz": [
+    -0.152,
+    -0.085,
+    -0.17
+   ],
+   "joint_xyz_after_apply_event_budget": [
+    0.1055,
+    -0.1247,
+    -0.37
+   ],
+   "joint_xyz_after_6s_more": [
+    -0.1116,
+    -0.085,
+    -0.3688
+   ],
+   "slot_centre_xy": [
+    -0.152,
+    0.085
+   ],
+   "tip_xy": [
+    -0.112,
+    0.085
+   ],
+   "tip_error_mm": 40.0,
+   "tip_z": -0.152,
+   "contacts": 12
+  },
+  "11": {
+   "commanded_joint_xyz": [
+    0.0,
+    -0.085,
+    -0.17
+   ],
+   "joint_xyz_after_apply_event_budget": [
+    -0.107,
+    -0.085,
+    -0.3685
+   ],
+   "joint_xyz_after_6s_more": [
+    -0.0168,
+    -0.085,
+    -0.3688
+   ],
+   "slot_centre_xy": [
+    0.0,
+    0.085
+   ],
+   "tip_xy": [
+    -0.017,
+    0.085
+   ],
+   "tip_error_mm": 17.0,
+   "tip_z": -0.152,
+   "contacts": 12
+  },
+  "12": {
+   "commanded_joint_xyz": [
+    0.152,
+    -0.085,
+    -0.17
+   ],
+   "joint_xyz_after_apply_event_budget": [
+    0.0003,
+    -0.085,
+    -0.3685
+   ],
+   "joint_xyz_after_6s_more": [
+    0.1271,
+    -0.085,
+    -0.3693
+   ],
+   "slot_centre_xy": [
+    0.152,
+    0.085
+   ],
+   "tip_xy": [
+    0.127,
+    0.085
+   ],
+   "tip_error_mm": 25.0,
+   "tip_z": -0.152,
+   "contacts": 16
+  }
+ },
+ "joint_xyz_for_two_wells_in_slot_2": {
+  "A1": [
+   0.0959,
+   -0.2149,
+   -0.3695
+  ],
+  "H12": [
+   0.0658,
+   -0.2398,
+   -0.3684
+  ]
+ },
+ "what_stops_the_gantry": {
+  "z_column_mass_kg": 0.2,
+  "SAFE_Z": {
+   "commanded": -0.02,
+   "j_z": -0.3544,
+   "tip_z": -0.1374,
+   "contacts": 16,
+   "actuator_force_N": 2674.9
+  },
+  "WORK_Z": {
+   "commanded": -0.17,
+   "j_z": -0.3693,
+   "tip_z": -0.1523,
+   "contacts": 16,
+   "actuator_force_N": 1594.6
+  },
+  "SAFE_Z, gravity off": {
+   "commanded": -0.02,
+   "j_z": -0.3618,
+   "contacts": 16
+  },
+  "contacts off": {
+   "commanded_xyz": [
+    -0.152,
+    -0.17,
+    -0.02
+   ],
+   "joint_xyz": [
+    -0.152,
+    -0.17,
+    -0.5169
+   ]
+  },
+  "contacts off, x after the 100-step budget": {
+   "commanded": 0.152,
+   "j_x": 0.1421
+  }
+ },
+ "liquid_bookkeeping": {
+  "pipette_uL_after_aspirating_5000": 5000.0,
+  "source_well_uL_after_aspirating_from_empty": 0.0,
+  "state_tracks_slot5_C7_uL": 100.0,
+  "well_geoms_that_change_colour": [
+   "w_A1",
+   "w_B1",
+   "w_C1",
+   "w_D1",
+   "w_E1",
+   "w_F1",
+   "w_G1",
+   "w_H1"
+  ]
+ },
+ "own_parser_on_opentrons_7_5": {
+  "log_lines": 27,
+  "aspirate_lines": 12,
+  "dispense_lines": 12,
+  "events_parsed": {
+   "pick": 1,
+   "drop": 1
+  },
+  "sample_line": "Aspirating 100.0 uL from A1 of reservoir on 1 at 92.86 uL/sec"
+ }
+}
diff --git a/research/expansion/tools/probe_mujoco_viz.py b/research/expansion/tools/probe_mujoco_viz.py
new file mode 100644
index 0000000..a894270
--- /dev/null
+++ b/research/expansion/tools/probe_mujoco_viz.py
@@ -0,0 +1,82 @@
+"""Empirical checks of opentrons-mujoco-viz (run with the interpreter that has it and mujoco installed).
+
+    python probe_mujoco_viz.py <opentrons_simulate> <protocol.py>  >  mujoco_viz_probe.json
+
+Measures rather than reads: which cameras render, whether the gantry reaches the position it is commanded to
+(by joint name, within the step budget apply_event uses and after six more seconds), what stops it (gravity and
+contacts toggled), what the liquid bookkeeping accepts, and what the package's own log parser gets from real
+Opentrons output.
+"""
+import json, subprocess, sys
+import mujoco
+from opentrons.visualization import WetLabScene, parse_simulate_output
+from opentrons.visualization.scene import SLOT_XY, WORK_Z, SAFE_Z
+
+out = {}
+with WetLabScene() as s:
+    m, d = s.model, s.data
+    q = lambda n: round(float(d.qpos[m.jnt_qposadr[mujoco.mj_name2id(m, mujoco.mjtObj.mjOBJ_JOINT, n)]]), 4)
+    tip = mujoco.mj_name2id(m, mujoco.mjtObj.mjOBJ_GEOM, "pip_tip")
+    out["model"] = dict(bodies=int(m.nbody), geoms=int(m.ngeom), joints=int(m.njnt), actuators=int(m.nu), cameras=int(m.ncam),
+                        meshes=int(m.nmesh), timestep_s=float(m.opt.timestep),
+                        geoms_with_collision_enabled=sum(1 for i in range(m.ngeom) if m.geom_contype[i] or m.geom_conaffinity[i]))
+    out["joint_ranges_m"] = {mujoco.mj_id2name(m, mujoco.mjtObj.mjOBJ_JOINT, i): [round(float(x), 3) for x in m.jnt_range[i]] for i in range(m.njnt)}
+    cams = {mujoco.mj_id2name(m, mujoco.mjtObj.mjOBJ_CAMERA, i): round(float(s.render(mujoco.mj_id2name(m, mujoco.mjtObj.mjOBJ_CAMERA, i)).mean()), 2)
+            for i in range(m.ncam)}
+    cams["free (fallback)"] = round(float(s.render("free").mean()), 2)
+    out["camera_mean_brightness_0_255"] = cams
+
+    rows = {}
+    for slot in range(1, 13):
+        sx, sy = SLOT_XY[slot]
+        s.move_to_slot(slot, work=True, steps=100)
+        budget = [q("j_x"), q("j_y"), q("j_z")]
+        for _ in range(3000):
+            mujoco.mj_step(m, d)
+        tx, ty, tz = (round(float(v), 3) for v in d.geom_xpos[tip])
+        rows[slot] = dict(commanded_joint_xyz=[round(sx, 3), round(sy - 0.17, 3), WORK_Z], joint_xyz_after_apply_event_budget=budget,
+                          joint_xyz_after_6s_more=[q("j_x"), q("j_y"), q("j_z")], slot_centre_xy=[sx, sy], tip_xy=[tx, ty],
+                          tip_error_mm=round(1000 * ((tx - sx) ** 2 + (ty - sy) ** 2) ** 0.5, 1), tip_z=tz, contacts=int(d.ncon))
+        s._move_to(d.ctrl[s._ax], d.ctrl[s._ay], SAFE_Z, steps=200)
+    out["gantry_by_slot"] = rows
+
+    wells = {}
+    for w in ("A1", "H12"):
+        s.apply_event({"type": "dispense", "volume": 10, "slot": 2, "well": w})
+        s.move_to_slot(2, work=True)
+        wells[w] = [q("j_x"), q("j_y"), q("j_z")]
+    out["joint_xyz_for_two_wells_in_slot_2"] = wells
+
+    zb = mujoco.mj_name2id(m, mujoco.mjtObj.mjOBJ_BODY, "z_column")
+    z = {"z_column_mass_kg": round(float(m.body_subtreemass[zb]), 3)}
+    for name, target in (("SAFE_Z", SAFE_Z), ("WORK_Z", WORK_Z)):
+        s._move_to(0.152, -0.17, target, steps=2000)
+        z[name] = dict(commanded=target, j_z=q("j_z"), tip_z=round(float(d.geom_xpos[tip][2]), 4), contacts=int(d.ncon),
+                       actuator_force_N=round(float(d.actuator_force[s._az]), 1))
+    m.opt.gravity[:] = 0
+    s._move_to(0.152, -0.17, SAFE_Z, steps=2000)
+    z["SAFE_Z, gravity off"] = dict(commanded=SAFE_Z, j_z=q("j_z"), contacts=int(d.ncon))
+    m.opt.disableflags |= mujoco.mjtDisableBit.mjDSBL_CONTACT
+    s._move_to(-0.152, -0.17, SAFE_Z, steps=2000)
+    z["contacts off"] = dict(commanded_xyz=[-0.152, -0.17, SAFE_Z], joint_xyz=[q("j_x"), q("j_y"), q("j_z")])
+    s._move_to(0.152, -0.17, SAFE_Z, steps=100)
+    z["contacts off, x after the 100-step budget"] = dict(commanded=0.152, j_x=q("j_x"))
+    out["what_stops_the_gantry"] = z
+
+    s.reset()
+    s.state.apply_event({"type": "aspirate", "volume": 5000, "slot": 3, "well": "A1"})
+    liquid = {"pipette_uL_after_aspirating_5000": s.state.pipette_vol, "source_well_uL_after_aspirating_from_empty": s.state.well_vol(3, "A1")}
+    s.state.apply_event({"type": "dispense", "volume": 100, "slot": 5, "well": "C7"})
+    liquid["state_tracks_slot5_C7_uL"] = s.state.well_vol(5, "C7")
+    liquid["well_geoms_that_change_colour"] = sorted(s._g_wells)
+    out["liquid_bookkeeping"] = liquid
+
+r = subprocess.run([sys.argv[1], sys.argv[2]], capture_output=True, text=True)
+lines = [l for l in r.stdout.splitlines() if l.strip()]
+kinds = {}
+for e in parse_simulate_output(r.stdout):
+    kinds[e["type"]] = kinds.get(e["type"], 0) + 1
+out["own_parser_on_opentrons_7_5"] = dict(log_lines=len(lines), aspirate_lines=sum("Aspirating" in l for l in lines),
+                                          dispense_lines=sum("Dispensing" in l for l in lines), events_parsed=kinds,
+                                          sample_line=next((l.strip() for l in lines if "Aspirating" in l), ""))
+print(json.dumps(out, indent=1))
diff --git a/research/expansion/tools/side_by_side.sh b/research/expansion/tools/side_by_side.sh
new file mode 100755
index 0000000..d722643
--- /dev/null
+++ b/research/expansion/tools/side_by_side.sh
@@ -0,0 +1,22 @@
+#!/bin/bash
+# Side-by-side PyLabRobot | MuJoCo for every task, both driven by the same oracle run log.
+S=/private/tmp/claude-501/-Users-nialloleary/36017c47-22e1-459a-91a2-0b4a5634f844/scratchpad
+L=$S/plr/logs; O=~/Desktop/Text2WetLab_SideBySide; mkdir -p $O $S/sbs
+cd $S/plr
+for t in colony-pcr-screening-hard ecoli-heat-shock-transformation-hard golden-gate-assembly-hard opentrons-rna-extraction-hard; do
+  [ -s cmp/$t.oracle.mp4 ] && continue
+  per=$(python3 -c "
+import json; e=json.load(open('logs/$t.oracle.json'))['events']
+n=sum(1 for x in e if x['kind'] in ('pick','drop','aspirate','dispense')); print(min(0.35, max(0.004, 55.0/max(n,1))))")
+  venv/bin/python replay_log.py logs/$t.oracle.json $t "Oracle (reference)" $per cmp/$t.oracle.mp4 cmp/$t.oracle.state.json < /dev/null 2>&1 | grep -v "^viewer" | tail -1
+done
+for f in $L/*.oracle.json; do
+  t=$(basename $f .oracle.json); p=$S/plr/cmp/$t.oracle.mp4
+  [ -s "$p" ] || { echo "no PLR video for $t"; continue; }
+  dur=$(ffprobe -v error -show_entries format=duration -of csv=p=0 $p)
+  (cd $S/mj && venv/bin/python mj_render.py $f $S/sbs/$t.mj.mp4 $dur "MuJoCo" < /dev/null > /dev/null 2>&1)
+  ffmpeg -nostdin -loglevel error -y -i $p -i $S/sbs/$t.mj.mp4 \
+    -filter_complex "[0:v]scale=960:540,tpad=stop_mode=clone:stop_duration=600[a];[1:v]scale=960:540,tpad=stop_mode=clone:stop_duration=600[b];[a][b]hstack=inputs=2[v]" \
+    -map "[v]" -t $dur -c:v libx264 -pix_fmt yuv420p -crf 23 $O/$t.mp4 && echo "side-by-side $t ($dur s)"
+done
+ls $O | wc -l