websum
Summarize web pages and YouTube videos with pluggable LLM backends. Ships with first-class support for Ollama (local) and OpenAI, plus an optional Gradio web UI.
Installation
# Library + CLI, with Ollama backend
pip install 'websum[ollama]'
# With OpenAI backend
pip install 'websum[openai]'
# With the Gradio web UI
pip install 'websum[ui,ollama]'
# Everything
pip install 'websum[all]'
Using uv:
uv add 'websum[ollama]'
Quickstart
Library
from websum import Summarizer, OllamaBackend
s = Summarizer(backend=OllamaBackend(model="llama3:instruct"))
print(s.summarize("https://cobanov.dev/haftalik-bulten/hafta-13"))
print(s.summarize("https://www.youtube.com/watch?v=4pOpQwiUVXc"))
print(s.translate("Hello world", target_language="Turkish"))
Swap the backend without touching anything else:
from websum import Summarizer, OpenAIBackend
s = Summarizer(backend=OpenAIBackend(model="gpt-4o-mini"))
CLI
# Summarize a web page or YouTube URL (auto-detected)
websum summarize https://example.com
# Use OpenAI instead of Ollama
websum summarize https://example.com --backend openai --model gpt-4o-mini
# Translate
websum translate "Hello world" --target-language Turkish
# Launch the Gradio UI
websum ui --port 7860
Run websum --help for the full command reference.
API overview
| Object | Purpose |
|---|---|
Summarizer |
High-level API. summarize(url), summarize_web(url), summarize_youtube(url), translate(text). |
SummarizerConfig |
Chunking and language settings. |
OllamaBackend, OpenAIBackend |
Built-in backends. Frozen dataclasses with .build(). |
LLMBackend (Protocol) |
Implement this to plug in any backend. |
BackendRegistry |
Map string names to backend classes (used by the CLI). |
All public names are re-exported from the top-level websum package and listed in __all__.
Custom backends
from dataclasses import dataclass
from websum import LLMBackend, Summarizer
@dataclass
class MyBackend:
def build(self):
from langchain_anthropic import ChatAnthropic
return ChatAnthropic(model="claude-3-5-sonnet-latest")
assert isinstance(MyBackend(), LLMBackend) # Protocol check
s = Summarizer(backend=MyBackend())
Docker
docker build -t websum .
docker run -p 7860:7860 websum
# Run when ollama is on the host
docker run --network host -p 7860:7860 websum
The image starts websum ui by default.
Migration from 0.1.x
The 0.1.x scripts under app/ (summarizer.py, translator.py, yt_summarizer.py, webui.py) are gone. Everything moved into the websum package with a typed, importable API.
| Before | After |
|---|---|
python app/summarizer.py -u URL |
websum summarize URL |
python app/webui.py |
websum ui |
from summarizer import setup_summarization_chain |
from websum import Summarizer |
Hardcoded ChatOllama |
OllamaBackend / OpenAIBackend / custom LLMBackend |
pip install -r requirements.txt |
pip install 'websum[ollama]' |
Development
git clone https://github.com/cobanov/websum
cd websum
uv sync --all-extras
uv run pre-commit install
uv run pytest
uv run ruff check .
uv run mypy src/websum
See CONTRIBUTING.md for the full guide.
License
MIT. See LICENSE.
Release files for websum 0.2.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| websum-0.2.0.tar.gz | 14.0 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| websum-0.2.0-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 27.3 kB
Release files / websum-0.2.0.tar.gz
| Download URL | websum-0.2.0.tar.gz |
|---|---|
| Size | 14.0 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
df527a9553c1f37012f85678fc3bfc89dbe3c7b6d6e60bbcc710eedb3b5c2b20
|
|
BLAKE2b-256 checksum How to use checksums |
688af7245642df724d08d4b584de7757da00a2a511aaaefe2a971f2e5d76d61f
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on May 14, 2026.
Transparency logRelease files / websum-0.2.0-py3-none-any.whl
| Download URL | websum-0.2.0-py3-none-any.whl |
|---|---|
| Size | 13.3 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
46c23197d69a2ee457039d51575874b5171b2b8f12a187ae56f4ba024363cc7d
|
|
BLAKE2b-256 checksum How to use checksums |
f303ff42ba931b2deedf031b381c43a37cd9d14c6b7617dc5ab43c9a09458404
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on May 14, 2026.
Transparency log