Skip to main content


Note For complete IPFSPy doc, please visit here

What is IPFSPy?

IPFSPy is a python library by Algovera to interact with IPFS and IPFS ecosystem such as the common pinning services. It is designed by data scientists for data scientists to interact with the IPFS ecosystem without leaving the comfort of python and jupyter notebook.

You can learn more about IPFS here.

IPFS is built using the go-lang or javascript. With IPFSPy, you can interact with IPFS using the exposed HTTP RPC API.

Ipfspy has two different implementations to interact with IPFS

  1. Access to most of what HTTP RPC API through ipfspy.ipfshttpapi

    With ipfspy.ipfshttpapi, you can either use local, infura or public nodes. In order to use local node, you will need a IPFS deamon running in the form of IPFS Desktop, IPFS Campanion or IPFS CLI. As an alternative, you can connect via the Infuraโ€™s dedicated IPFS gateway.

  2. fsspec-like implementation through ipfspy.ipfspec

    What is ipfsspec?

    ipfsspec is the fsspec-like implementation for IPFS. Through ipfspy.ipfsspec, you get access to both local and infura nodes but the public node is read-only. Local node offers both read and write. In order to use local node, you will need a local node running locally through IPFS Desktop or IPFS Companion or the IPFS CLI.

    Advantages of fsspec-like implementations

    • many data science libraries like dask, pandas and xarray uses fsspec for path and file handling
    • provides an uniform APIs

    Since many data science use fsspec, by implementing ipfsspec, fsspec supports IPFS.


You can install the library simply through

pip install ipfspy

If you plan to contribute to this library, 1. clone the repo locally git clone

  1. Install an editable version of the repo pip install -e .[dev]

  2. submit PR

How to use


Extend the IPFSApi

api = IPFSApi()
Changed to local node

Change to other nodes like below. You can choose one of local infura or public

api.change_gateway_type = 'infura'
Changed to infura node
api.change_gateway_type = 'local'
Changed to local node

To add a file to IPFS,

res, obj = api.add_items('output/test.txt'); obj
[{'Name': 'test.txt', 'Bytes': 20},
{'Name': 'test.txt',
'Hash': 'QmbwFKzLj9m2qwBFYTVrBotf4QujvzTb1GBV9wFcNPMctm',
'Size': '28'}]

To retrieve a file from IPFS

response, content = api.cat_items(cid='QmbwFKzLj9m2qwBFYTVrBotf4QujvzTb1GBV9wFcNPMctm'); content
"\n'''\nFirst file\n'''\n"


import fsspec, io, os, asyncio
from ipfspy.ipfsspec.asyn import AsyncIPFSFileSystem
from fsspec import register_implementation

Register the ipfsspec to fsspec

register_implementation(AsyncIPFSFileSystem.protocol, AsyncIPFSFileSystem)
class fs:
ipfs = fsspec.filesystem("ipfs")
file = fsspec.filesystem("file")
Changed to local node

Like ipfspy.ipfshttpapi, you can change the node as such

fs.ipfs.change_gateway_type = 'public'
Changed to public node
fs.ipfs.change_gateway_type = 'local'
Changed to local node

Add a file to IPFS using fsspec-like api

put_file_res = fs.ipfs.put(path='output/test.txt', rpath='/test_put_file'); put_file_res

Retrieve a file from IPFs using fsspec-like api'QmbwFKzLj9m2qwBFYTVrBotf4QujvzTb1GBV9wFcNPMctm')
b"\n'''\nFirst file\n'''\n"