NEW — Full Python with import math, classes, and stdlib

Your Python runs
47× faster here

Python with import math and numpy compiled to native Wasm. Every execution produces a SHA-256 receipt. Self-hosted for free.

.cell (Rust)
<1ms
67–133× faster
.cell (Full Python)
~72ms
7× faster
E2B (warm)
~100ms*
Pre-booted VM
E2B (cold)
~500ms*
Firecracker VM
*E2B numbers cited from public sources, not independently measured.

Why .cell wins

Python compiled to native Wasm. No Docker. No VMs. No kernel.

47× Faster Python

Python with import math and numpy compiles to native Wasm. No interpreter overhead. 18 math functions, 15 numpy operations.

🔐

Cryptographic Receipts

Every execution returns SHA-256 hashes of code + result. Patent-grade proof. Regulatory audit trails. Reproducible science.

🛡️

Hardened Security

Dangerous builtins blocked at the compiler level. No filesystem. No network. Fuel-metered execution. 52 adversarial attack vectors tested and blocked.

🔌

MCP Native

Drop-in MCP server for Claude, Cursor, Windsurf. One config line to add verified code execution to any AI agent.

🏠

Self-Hosted for Free

Single Rust binary. No Docker, no Kubernetes. Your metal, your data. E2B charges enterprise pricing for self-hosted.

🔬

Research-Grade

Built for R&D. Arrhenius kinetics, parameter sweeps, Monte Carlo. Every result cryptographically verified for patent filings.

Three lines of code

Create a cell. Execute code. Get a receipt.

curl — Create & Execute
# Create a JavaScript cell curl -X POST http://cell.synapserun.dev/v1/cells \ -H 'Authorization: Bearer YOUR_KEY' \ -H 'Content-Type: application/json' \ -d '{"template":"javascript"}' # Execute code curl -X POST ".../cells/{id}/exec" \ -d '{"code":"console.log(42 * 42)"}'
Response — 0.62ms with receipt
{ "stdout": "1764", "stderr": "", "exit_code": 0, "latency_ms": 0.62, "receipt": { "execution_id": "1656c200-...", "code_hash": "f2f248f0...", "result_hash": "02787652..." } }

Try it live

Execute on our AX102 server. Real Wasm sandbox. Real cryptographic receipts.

Ready — runs at native Wasm speed
OUTPUT
Click "Execute" to run your code on AX102...

.cell vs E2B

Side-by-side. Honest methodology.

CapabilityE2B.cellWinner
Cold start~500ms (cited*)<2ms67–133× .cell
Full PythonCPython on Linux VMCPython-WASI (~72ms)7× .cell (vs cold)
Compute PythonCPython (interpreted)Compiled to Wasm (<1ms)67–133× .cell
math / numpy✅ (C extensions)✅ via CPython-WASIParity
JavaScript✅ ES2023 (QuickJS)Parity
Execution receipts✅ SHA-256.cell only
Deterministic✅ Bit-identical.cell only
Self-hostedEnterprise $$$✅ Free.cell
Security auditUnknown52 adversarial attacks blocked.cell
PricingUsage-basedUsage-basedSame price, faster
HTTP/Networking✅ Full networking❌ Coming soonE2B (for now)
Persistent sessions❌ Coming soonE2B (for now)

*E2B cold-start numbers cited from public documentation. Not independently benchmarked.

Start executing in 30 seconds

One curl command. No signup. No credit card. No Docker.

# Python SDK pip install synapse-cell # 3 lines to execute code: from synapse_cell import Cell cell = Cell() result = cell.run("print(42 * 42)")
# MCP Server (Claude Desktop) { "mcpServers": { "cell": { "command": "npx", "args": ["@synapse/mcp-server"] } } }
Request API Key 🔐 Verify Receipt 📖 Read the Blog