Tools · Cheminformatics & Structure

MolGen

Generate novel, chemically valid analogs of a seed molecule with a SELFIES molecular language model

On this page

Prices, workflows, and method papers Open in the app

Run it from the API

Submit with Submit a job and the job_type below. Price it first with Estimate job reservation cost: submitting reserves that amount from your wallet, and the charge settles at the actual runtime.

MolGen `molgen`

Job type: molgen
Hardware: t4l4a10 (default)l40sa100-40gba100-80gbh100h200b200
Typical runtime: 10 min on Nvidia A10

Payload

Payload fields
Field	Type	Description
`mode`required	`string`	Default: `"generate"`One of: `"generate", "sample"`
`input_data[]`	`(string \| object)[]`	Limits: `max items 1000`
`input_format`	`string`	Default: `"smiles"`One of: `"smiles", "sdf", "molblock", "inchi"`
`num_sequences`	`integer`	Sequences to generate per seed (at most 100 when seeds are present, with len(input_data) × num_sequences <= 10000), or in total (at most 10000) for unconditional sample mode with an empty input_data. Default: `10`Limits: `≥ 1, ≤ 10000`
`max_length`	`integer`	Maximum SELFIES token length per generated sequence. Default: `64`Limits: `≥ 8, ≤ 128`
`num_beams`	`integer`	Default: `5`Limits: `≥ 1, ≤ 10`
`temperature`	`number`	Default: `1`Limits: `≥ 0.1, ≤ 2`
`top_p`	`number`	Default: `0.95`Limits: `≥ 0, ≤ 1`
`top_k`	`integer`	0 disables top-k filtering. Default: `0`Limits: `≥ 0, ≤ 1000`
`repetition_penalty`	`number`	Default: `1`Limits: `≥ 1, ≤ 2`
`exclude_inputs`	`boolean`	Drop generated molecules that match an input seed (canonical comparison). Default: `false`
`max_output_molecules`	`integer`	Upper bound on deduplicated molecules emitted. Default: `1000`Limits: `≥ 1, ≤ 10000`

Example

from cognichem_client import CogniChem

client = CogniChem.from_env()  # reads COGNICHEM_API_KEY
payload = {
    "mode": "generate",
    "input_data": ["c1ccccc1"],
    "input_format": "smiles",
    "num_sequences": 4,
    "max_length": 64,
    "num_beams": 5,
    "temperature": 1,
    "top_p": 0.95,
    "top_k": 0,
    "repetition_penalty": 1,
    "exclude_inputs": True,
    "max_output_molecules": 1000,
}

estimate = client.jobs.estimate(job_type="molgen", payload=payload, resource="a10")
print(f"Reserves ${estimate.cost:.2f}")

job = client.jobs.submit(
    job_name="my-molgen-run",
    job_type="molgen",
    payload=payload,
    resource="a10",
)
status = client.jobs.wait(job.process_id)
if status.status == "completed":
    client.jobs.result(job.process_id, save_path=".")

curl -X POST "https://api.cognichem.com/api/v1/jobs/submit" \
  -H "X-Api-Key: $COGNICHEM_API_KEY" \
  -H "Idempotency-Key: $(uuidgen)" \
  -H "Content-Type: application/json" \
  -d '{
  "job_name": "my-molgen-run",
  "job_type": "molgen",
  "payload": {
    "mode": "generate",
    "input_data": [
      "c1ccccc1"
    ],
    "input_format": "smiles",
    "num_sequences": 4,
    "max_length": 64,
    "num_beams": 5,
    "temperature": 1,
    "top_p": 0.95,
    "top_k": 0,
    "repetition_penalty": 1,
    "exclude_inputs": true,
    "max_output_molecules": 1000
  },
  "resource": "a10"
}'

{
  "mode": "generate",
  "input_data": [
    "c1ccccc1"
  ],
  "input_format": "smiles",
  "num_sequences": 4,
  "max_length": 64,
  "num_beams": 5,
  "temperature": 1,
  "top_p": 0.95,
  "top_k": 0,
  "repetition_penalty": 1,
  "exclude_inputs": true,
  "max_output_molecules": 1000
}

Sample data from the job catalog; long values are shortened here. Each job_name must be unique among your jobs.

Workflow inputs

Molecules (list) · optionalSMILES, SDF, MOLBLOCK, INCHI

Workflow outputs

ArchiveZIP
TableCSV
MoleculesSMILES

Run it from the API

MolGen molgen

MolGen `molgen`