Data

Keep artifacts and source close to the agents that use them.

Create private buckets and Git repositories in the same Tarka account as your agents, jobs and models, from the console, the CLI or your coding agent.

Open console Read docs

S3-compatible
Versioned buckets
Private Git
Git LFS

# a private, versioned bucket
tarka buckets create reports
 
# a key for that bucket only, saved as AWS_* variables
tarka buckets credentials create reports > .env.s3
 
# a private Git repository with large file support
tarka git create support-agent

Object storage

Private objects through an S3-compatible API.

Each key opens one bucket. Point the AWS CLI, boto3 or any S3 SDK at it for datasets, documents and the files your agents produce, and the bucket keeps earlier versions of every object.

Object storage docs

# the key the hero saved in .env.s3
set -a; . ./.env.s3; set +a
 
# then the AWS CLI works as it does with S3
aws s3 cp invoice.pdf "s3://$TARKA_S3_BUCKET/invoices/"
aws s3 ls "s3://$TARKA_S3_BUCKET/invoices/"

import os, boto3  # pip install boto3 python-dotenv
from dotenv import load_dotenv
 
load_dotenv(".env.s3")  # the key the hero saved
s3 = boto3.client("s3")  # endpoint included
s3.upload_file("invoice.pdf", os.environ["TARKA_S3_BUCKET"],
               "invoices/invoice.pdf")

Hosted Git

Private source with large file support.

Clone over SSH or HTTPS, push from your coding agent or CI, and keep large assets in Git LFS, next to the agents and jobs that build from them.

Hosted Git docs

# your hosted Git account, once; add an SSH key to it
tarka git setup
# clone with the URL tarka git create printed
git clone ssh://git@git.tarka.rest:2222/you/support-agent.git
cd support-agent
 
# large files go to Git LFS
git lfs track "*.pdf"
git add .gitattributes help-center/
git commit -m "Add the help center" && git push

Create an account to set up buckets and repositories.

Create an account Read the data docs