[SYS_OK] MEA_0 ACTIVE -- 862 Hz mean spike rate -- donor 924A consent verified -- 37.0 C -- rack draw 914 W -- 0.85 Gop/W -- MEA_1 ACTIVE -- 743 Hz -- donor 482B consent verified -- 37.1 C -- perfusion nominal --
OrganoidGrid
Platform profile · FinalSpark

FinalSpark Neuroplatform

The Neuroplatform is biocomputing without an incubator: living human brain organoids kept alive in a Swiss lab, reachable over the internet through an API so you can stimulate them and read their activity from anywhere.

Its real innovation is not a faster organoid. It is the idea that the hardest part of biocomputing, keeping the tissue alive, can be centralized and rented. That turns a field gated by cell-culture expertise into something a software researcher can touch. It also concentrates all the biology, and all the ethics, in one operator's hands.

Translucent brain organoids on microelectrode arrays in a laboratory rack under a faint green glow, against a black background.
The Neuroplatform centralizes the life support and exposes the tissue over an API. Imaging is illustrative.

What is the Neuroplatform?

FinalSpark runs clusters of three-dimensional human brain organoids, each sitting on a microelectrode array inside a perfusion chamber, and exposes them through REST and WebSocket interfaces.4 A researcher sends a stimulation vector as a request and streams back the recorded field potentials. The organoids and their life support never leave the lab. This is the Wetware-as-a-Service model in practice.

3Dcerebral organoids
~25 kHzrecording stream
REST + WSremote interface
37.0 Cchamber setpoint

Why keeping a 3D organoid alive is the hard part

A planar culture is a thin film; a brain organoid is a sphere up to a couple of millimeters across, and that geometry is the whole engineering problem. Oxygen and glucose have to diffuse inward to the core while waste diffuses out, and past a certain radius the center starves and dies, the necrotic-core problem that has shaped organoid work since the first cerebral organoids were grown.11 Continuous microfluidic perfusion, real-time pH and oxygen sensing, and tight thermal control are what hold a culture in the viable band.

Diffusion gradient across a perfused brain organoid A cross-section from the perfused surface of an organoid inward to its core, showing oxygen and nutrient supply falling and necrosis risk rising toward the center. surface -> core Perfused surface high O2 / glucose, electrodes here Outer shell mature, active neurons Transition zone falling oxygen Core necrosis risk without perfusion
Across the radius of a 3D organoid, oxygen and nutrient availability fall toward the core while waste accumulates. Perfusion engineering exists to keep the interior inside the viable band.

How the platform writes to the tissue

Writing happens two ways. Electrically, charge-balanced biphasic pulses on the array drive localized firing and, with high-frequency bursts, push synapses toward long-term potentiation. Chemically, the microfluidics can inject micro-doses of neuromodulators such as dopamine into specific wells. Pairing a desired outcome with a dopamine pulse makes the affected synapses more plastic, a crude but real reward signal that lets an operator nudge the network's self-organization rather than program it.

The remote programming model

From the developer's side it looks like any networked service. You send a JSON stimulation vector (target chamber, electrode IDs, pulse parameters) to a REST endpoint; a WebSocket streams the multichannel recording back in near real time for local filtering and spike-sorting. The cost of that convenience is latency: the network round trip rules out the sub-millisecond closed loops a local unit allows, so the Neuroplatform suits exploration, plasticity studies, and screening better than tight real-time control.

Signal acquisition and feedback pipeline A left-to-right chain of processing stages from the electrode array through amplification, digitization, spike sorting and decoding, then back to the stimulator. Client JSON vector REST queue stim Organoid living tissue ADC ~25 kHz WebSocket stream back
The Wetware-as-a-Service request loop. A client sends a stimulation vector to the REST endpoint; the platform delivers it to a living organoid and streams the digitized response back over a WebSocket. Network latency, not the tissue, sets the loop time.

Access and what you get

FinalSpark offers subscription access under research and educational plans, with commercial terms by consultation. A subscription allocates compute slots on specific organoid chambers, an API key and Python SDK, and a dashboard exposing the physiological telemetry (temperature, pH, flow) of your allocated tissue, which is the right thing to demand: if you are computing on something alive, you should be able to watch its vital signs.

Frequently asked questions

What is the FinalSpark Neuroplatform?

A remotely accessible bio-cloud of living human brain organoids on microelectrode arrays, reachable over REST and WebSocket APIs so researchers can stimulate and record the tissue without local lab infrastructure.

How is it different from the Cortical Labs CL1?

The CL1 is local hardware you own and operate; the Neuroplatform is a remote service you rent. Local gives lower latency for real-time control; remote removes the burden of running a culture lab.

Why do 3D organoids need such elaborate life support?

Because a sphere of tissue must get oxygen and nutrients to its core by diffusion. Past a certain size the center dies without continuous perfusion, so the platform runs constant microfluidic flow and environmental control.

Can it do real-time closed-loop control?

Less well than local hardware. Network latency sets the loop time, so the platform suits exploration, plasticity studies, and screening more than sub-millisecond control.

References

  1. Jordan FD, et al. Open and remotely accessible Neuroplatform for research in wetware computing. Frontiers in Artificial Intelligence. 2024;7:1376042. doi:10.3389/frai.2024.1376042. Accessed 2026-06-12.
  2. Lancaster MA, et al. Cerebral organoids model human brain development and microcephaly. Nature. 2013;501(7467):373-379. doi:10.1038/nature12517. Accessed 2026-06-12.