mfsig.com
Live demo

Try mfsig.com

Submit a SMILES, get a σ-profile back. This UI calls a self-hosted molforge-backend (see backend/README.md). For production keys, contact your admin.

Generate a .mfsig — preview free, buy to download

Paste a structure, preview every channel free, then buy the signed file at the grade your pipeline needs. Paid jobs run on an on-demand GPU pod — you only pay for compute you keep.

1 · Pick a sample, see a result

Reference σ-profiles for the 100-drug Ground-Truth cohort are already minted. Open any one in the 3D viewer — molecule, σ-surface, audit pack, and σ-profile chart, all in your browser, no setup.

Open the 3D viewer →

2 · Convert your own legacy file

Drop a Turbomole / COSMOtherm / Gaussian / NwChem / ORCA / xtb σ-profile file into the converter. Get back a .mfsig.json with a SHA-256 audit trail you can verify forever. Open-source, no API key.

Open the converter →
Production flow

Submit a SMILES via API

Three calls, two optional. Webhook delivery means your worker never has to poll.

Step 1 · submit
curl -X POST https://api.mfsig.com/v1/sigma_profile \
  -H "Authorization: Bearer $KEY" \
  -d '{
    "smiles":  "CC(=O)Oc1ccccc1C(=O)O",
    "tier":    "platinum",
    "solvent": "water",
    "webhook_url": "https://you.app/cb"
  }'

Returns job_id + ETA in seconds.

Step 2 · we run
  • • Conformer search + pre-optimisation
  • • Geometry refinement
  • • Quantum single-point on the relaxed structure
  • • Analytical molecular surface (Phase 1+2+3)
  • • σ-profile + σ-moments + HB-split
  • • Ed25519 sign + audit chain

All 6 audit gates checked per drug. Re-run on us if any fails.

Step 3 · deliver
{
  "mfsig_url":"https://mfsig.com/d/job_8K2.mfsig.json",
  "sha256":   "e3b0c44298fc1c14...",
  "gates": {
    "scf_converged":      true,
    "sigma_sum_positive": true,
    "sigma_segments":     true,
    "geometry_ok":        true,
    "mfsig_verified":     true,
    "positions_populated":true
  }
}

Open the URL in the 3D viewer or download.

What lands

The .mfsig.json shape

Every delivery is a single file. Five pillars, one SHA-256 signature.

aspirin.mfsig.json (truncated)json
{
  "mfsig_version": "2.0-apex",
  "audit_and_trust": {
    "molecule_name":   "aspirin",
    "sha256_checksum": "e3b0c44298fc1c14...",
    "timestamp_utc":   "2026-05-12T08:52Z",
    "uuid":            "d42b8a5d-2211-4f51-825b-…",
    "reproducibility": { "molforge_version": "0.28.0", ... }
  },
  "chemistry_and_geometry": {
    "smiles":          "CC(=O)Oc1ccccc1C(=O)O",
    "inchi_key_14":    "BSYNRYMUTXBXSQ",
    "atomic_numbers":  [6, 6, 8, 8, ..., 1],
    "positions_aa":    [[x1, y1, z1], ..., [xN, yN, zN]],   // relaxed geometry
    "bonds":           [[0,1,1], [1,2,2], ...]
  },
  "quantum_and_thermodynamics": {
    "methodology": { "functional":"B3LYP", "basis":"def2-SVP", "solvent":"water" },
    "results": {
      "sigma_profile_51":  [...],   "sigma_profile_101": [...],   "sigma_profile_201": [...],
      "sigma_moments":     { "mean":..., "variance":..., "sigma_donor_mass":..., ... },
      "sigma_surface_segments": [ { "position":[...], "sigma":..., "area":..., ... } ],
      "atomic_charges_mulliken": [...],
      "total_energy_hartree":   -648.236...,
      "g_polar_hartree":        -0.015...,
      "geometry_provenance":    { "xtb_relax_converged": true, ... },
      "compute_provenance":     { "scc_wall_s":..., "gpu_backend_used":"gpu", ... }
    }
  },
  "ai_intelligence_layer": { "is_differentiable": true, "cavity_latent_vector_512d": null }
}

Every numeric field is needed to reproduce the σ-profile bit-exactly on the same hardware. The sha256_checksum is over the canonical state and validates via verify_integrity().