Check out the HTTP Client reference for the full command reference.

Note: The ipfsapi PIP package and Python module have both been renamed to ipfshttpclient! See the relevant section of the CHANGELOG for details. There is also a ipfsApi library from which this library originated that is completely unmaintained and does not work with any recent go-IPFS version.

Note: This library occasionally has to change to stay compatible with the IPFS HTTP API. Currently, this library is tested against go-ipfs v0.8.0. We strive to support the last 5 releases of go-IPFS at any given time; go-IPFS v0.5.0 therefore being to oldest supported version at this time.

• Install
• Usage
• Documentation
  • Migrating from 0.4.x to 0.6.0
• Featured Projects
• Contributing
  • Easy Tasks
  • Bug reports
  • Setting up a local development environment
  • Pull requests
  • Chat with Us (IRC/Matrix)
• License

Install with pip:

pip install ipfshttpclient

# Clone the source repository
git clone https://github.com/ipfs/py-ipfs-http-client.git
cd py-ipfs-http-client

# Link ipfs-api-client into your Python Path
flit install --pth-file

Basic use-case (requires a running instance of IPFS daemon):

>>> import ipfshttpclient
>>> client = ipfshttpclient.connect()  # Connects to: /dns/localhost/tcp/5001/http
>>> res = client.add('test.txt')
>>> res
{'Hash': 'QmWxS5aNTFEc9XbMX1ASvLET1zrqEaTssqt33rVZQCQb22', 'Name': 'test.txt'}
>>> client.cat(res['Hash'])
'fdsafkljdskafjaksdjf\n'

Please note: You should specify the address for an IPFS API server, using the address of a gateway (such as the public ipfs.io one at /dns/ipfs.io/tcp/443/https) will only give you extremely limited access and may not work at all. If you are only interested in downloading IPFS content through public gateway servers then this library is unlikely of being of much help.

For real-world scripts you can reuse TCP connections using a context manager or manually closing the session after use:

import ipfshttpclient

# Share TCP connections using a context manager
with ipfshttpclient.connect() as client:
	hash = client.add('test.txt')['Hash']
	print(client.stat(hash))

# Share TCP connections until the client session is closed
class SomeObject:
	def __init__(self):
		self._client = ipfshttpclient.connect(session=True)

	def do_something(self):
		hash = self._client.add('test.txt')['Hash']
		print(self._client.stat(hash))

	def close(self):  # Call this when your done
		self._client.close()

Administrative functions:

>>> client.id()
{'Addresses': ['/ip4/127.0.0.1/tcp/4001/ipfs/QmS2C4MjZsv2iP1UDMMLCYqJ4WeJw8n3vXx1VKxW1UbqHS',
               '/ip6/::1/tcp/4001/ipfs/QmS2C4MjZsv2iP1UDMMLCYqJ4WeJw8n3vXx1VKxW1UbqHS'],
 'AgentVersion': 'go-ipfs/0.4.10',
 'ID': 'QmS2C4MjZsv2iP1UDMMLCYqJ4WeJw8n3vXx1VKxW1UbqHS',
 'ProtocolVersion': 'ipfs/0.1.0',
 'PublicKey': 'CAASpgIwgg ... 3FcjAgMBAAE='}

Pass in API options:

>>> client.pin.ls(type='all')
{'Keys': {'QmNMELyizsfFdNZW3yKTi1SE2pErifwDTXx6vvQBfwcJbU': {'Count': 1,
                                                             'Type': 'indirect'},
          'QmNQ1h6o1xJARvYzwmySPsuv9L5XfzS4WTvJSTAWwYRSd8': {'Count': 1,
                                                             'Type': 'indirect'},
          …

Add a directory and match against a filename pattern:

>>> client.add('photos', pattern='*.jpg')
[{'Hash': 'QmcqBstfu5AWpXUqbucwimmWdJbu89qqYmE3WXVktvaXhX',
  'Name': 'photos/photo1.jpg'},
 {'Hash': 'QmSbmgg7kYwkSNzGLvWELnw1KthvTAMszN5TNg3XQ799Fu',
  'Name': 'photos/photo2.jpg'},
 {'Hash': 'Qma6K85PJ8dN3qWjxgsDNaMjWjTNy8ygUWXH2kfoq9bVxH',
  'Name': 'photos/photo3.jpg'}]

Or add a directory recursively:

>>> client.add('fake_dir', recursive=True)
[{'Hash': 'QmQcCtMgLVwvMQGu6mvsRYLjwqrZJcYtH4mboM9urWW9vX',
  'Name': 'fake_dir/fsdfgh'},
 {'Hash': 'QmNuvmuFeeWWpxjCQwLkHshr8iqhGLWXFzSGzafBeawTTZ',
  'Name': 'fake_dir/test2/llllg'},
 {'Hash': 'QmX1dd5DtkgoiYRKaPQPTCtXArUu4jEZ62rJBUcd5WhxAZ',
  'Name': 'fake_dir/test2'},
 {'Hash': 'Qmenzb5J4fR9c69BbpbBhPTSp2Snjthu2hKPWGPPJUHb9M',
  'Name': 'fake_dir'}]

This module also contains some helper functions for adding strings and JSON to IPFS:

>>> lst = [1, 77, 'lol']
>>> client.add_json(lst)
'QmQ4R5cCUYBWiJpNL7mFe4LDrwD6qBr5Re17BoRAY9VNpd'
>>> client.get_json(_)
[1, 77, 'lol']

Use an IPFS server with basic auth (replace username and password with real creds):

>>> import ipfshttpclient
>>> client = ipfshttpclient.connect('/dns/ipfs-api.example.com/tcp/443/https', auth=("username", "password"))

Pass custom headers to the IPFS daemon with each request:

>>> import ipfshttpclient
>>> headers = {"CustomHeader": "foobar"}
>>> client = ipfshttpclient.connect('/dns/ipfs-api.example.com/tcp/443/https', headers=headers)

Connect to the IPFS daemon using a Unix domain socket (plain HTTP only):

>>> import ipfshttpclient
>>> client = ipfshttpclient.connect("/unix/run/ipfs/ipfs.sock")

Documentation (currently mostly API documentation unfortunately) is available on IPFS:

https://ipfs.io/ipns/12D3KooWEqnTdgqHnkkwarSrJjeMP2ZJiADWLYADaNvUb6SQNyPF/docs/

The ipfs command-line Client documentation may also be useful in some cases.

Please see the CHANGELOG for the minor breaking changes between these releases.

Projects that currently use py-ipfs-http-client. If your project isn't here, feel free to submit a PR to add it!

• InterPlanetary Wayback interfaces web archive (WARC) files for distributed indexing and replay using IPFS.

Over time many smaller day-to-day tasks have piled up (mostly supporting some newer APIs). If want to help out without getting too involved picking up one of tasks of our help wanted issue list would go a long way towards making this library more feature-complete. 👍

You can submit bug reports using the GitHub issue tracker.

1. Follow the instructions in the IPFS documentation to install go-IPFS into your ${PATH}:
   https://docs.ipfs.io/install/command-line/
2. Follow the instructions in the (Python) tox documentation to install the tox Python environment runner:
   https://tox.readthedocs.io/en/latest/install.html
3. Clone the GIT repository if you haven't already:
   git clone https://github.com/ipfs-shipyard/py-ipfs-http-client.git

If you want to you can also make your local development version of py-ipfs-http-client available to your Python environment by installing flit and running flit install --pth-file from the repository root.

Please see the next section for how to run tests and contribute code back into the project.

Pull requests are welcome. Before submitting a new pull request, please make sure that your code passes both the code formatting (PEP-8 with tab indentation) and typing (PEP-484 using mypy) checks:

$ tox -e styleck -e typeck

As well as the unit tests:

$ tox -e py3 -e py3-httpx

If you are unsure, don't hesitate to just submit your code, and a human will take a look. 🙂

If you can, Please make sure to include new unit tests for new features or changes in behavior. We aim to bring coverage to 100% at some point.

You can arrange for the code style and typing tests to be run automatically before each commit by installing the GIT pre-commit hook:

$ ./tools/pre-commit --install

You can find us on #py-ipfs on chat.freenode.org or in our Matrix chat room. Join us if you have any suggestions, questions or if you just want to discuss IPFS and Python in general.

Please note that the channel is not monitored all the time and hence you may only receive a reply to your message later that day. Using Matrix makes it easier to stay connected in the background, so please prefer the Matrix option or use an IRC bouncer.

This code is distributed under the terms of the MIT license. Details can be found in the file LICENSE in this repository.