API Documentation
    Setup

    Overview

    Overview and rate limits

    API Key

    Setup and limits

    MCP Server

    Connect AI assistants to Tamarind tools

    CLI

    Run Tamarind from your terminal, scripts, or agents
    GET

    View available tools and configurations

    Submit Jobs
    POST

    Submit individual jobs

    POST

    Submit batch jobs

    POST

    Validate job settings without submitting

    Finetuning
    GET

    List available finetuned models

    Job Management
    GET

    List and inspect jobs

    POST

    Download job results

    DELETE

    Delete jobs and associated data

    POST

    Stop running or queued jobs

    File Management
    PUT

    Upload files to your account

    GET

    List and organize files

    DELETE

    Remove files from your account

    Pipelines
    POST

    Run saved pipelines

    POST

    Create new pipelines

Submit Single Jobs

POST/api/submit-job

The POST API endpoint /submit-job allows users to submit jobs to Tamarind. These jobs are visible in the same queue as jobs submitted through the UI. You may choose your job type and settings below to configure your job.

Options for inputting file fields (ex. .pdb, .sdf, etc.):

  • Use the /upload endpoint and upload your file before submitting
  • Use the path of a previous Tamarind job, in the format JobName/path/to/file.ext
  • Submit the contents of your file directly

Select tool:

Settings

Protein amino acid sequence, use : to separate chains for multimers

Options: ["1", "2", "3", "4", "5"]

Default: 5

Number of Models: Number of models to be used, each generating a different prediction

Default: 3

Number of recycles: Number of times to recycle outputs back through structure prediction process for refined results

Default: 0

Number of models to relax: Number of models to perform amber relaxation for more accurate side chain predictions

Default: True

Use Multiple Sequence Alignment: When MSAs are disabled, AlphaFold will run in single sequence mode.

Options: ["paired", "unpaired", "unpaired_paired"]

Pair Mode: "unpaired_paired" = pair sequences from same species + unpaired MSA, "unpaired" = separate MSA for each chain, "paired" - only use paired sequences.

Options: ["uniref", "swissprot", "uniref+swissprot"]

Default: uniref

MSA Database: Retired. MSAs are always built against UniRef30 plus the ColabFold environmental database. Legacy values are accepted and ignored.

Options: ["pdb100", "custom", "none"]

Default: pdb100

Template Mode: Choose which template mode to use for your prediction

Custom Template File: One structural template for the prediction (.cif or .pdb). Wire a single structure, or upload one file.

Initial Guess Structure: Optional PDB/CIF whose atom positions seed the prediction instead of a random starting model (ColabFold --initial-guess). Useful for refining a known or designed structure.

Random seed: Random seed to be used in structure prediction

Options: ["508:2048", "512:1024", "256:512", "128:256", "64:128", "32:64", "16:32"]

Max MSA: Max # Clusters : Max # Extra Sequences - decrease max_msa to increase uncertainty

Options: ["0.0", "0.5", "1.0"]

Recycle Early Stop Tolerance: Run recycles until distance between recycles is within a given tolerance (0 = never stop early)

Default: False

Score with IPSAE: Use IPSAE scoring function for interprotein interactions

Default: False

Options: ["auto", "alphafold2_ptm", "alphafold2_multimer_v1", "alphafold2_multimer_v2", "alphafold2_multimer_v3", "deepfold_v1", "alphafold2"]

Default: auto

Model Type: Model type to be used. If auto selected, will use alphafold2_ptm for monomer prediction and alphafold2_multimer_v3 for complex prediction (recommended canonical weights). Any of the mode_types can be used (regardless if input is monomer or complex)

HTTP Response Status Codes
Status codeDescription
200Job successfully submitted
400Bad request
403Forbidden - organization or team budget exceeded
500Internal server error
1import requests
2
3api_key = "***************"
4headers = {'x-api-key': api_key}
5base_url = "https://app.tamarind.bio/api/"
6
7params = {
8  "jobName": "myJobName",
9  "type": "alphafold",
10  "settings": {
11    "sequence": "MALKSLVLLSLLVLVLLLVRVQPSLGKETAAAKFERQHMDSSTSAASSSNYCNQMMKSRNLTKDRCKPVNTFVHESLADVQAVCSQKNVACKNGQTNCYQSYSTMSITDCRETGSSKYPNCAYKTTQANKHIIVACEGNPYVPVHFDASV"
12  }
13}
14response = requests.post(base_url + "submit-job", headers=headers, json=params)
15print(response.text)
Response Format
myJobName submitted to queue.
Optional Parameters:

projectTag

= "proj_..."

Assign the job to a project by its ProjectId