fastapi-hotwire
fastapi-hotwire brings Hotwire's Turbo protocol to FastAPI. Render targeted DOM updates with <turbo-stream> responses, return individual Jinja blocks for <turbo-frame> requests, queue session-backed flash messages, and validate forms in place — all from your existing FastAPI handlers, with no JSON layer or client-side framework. Ships with pytest helpers and a Protocol-based design so you can swap template engines or session backends.
Install
pip install fastapi-hotwire
# or
uv add fastapi-hotwire
For the Pydantic-backed validation-error stream:
pip install "fastapi-hotwire[forms]"
Quickstart
from fastapi import FastAPI, Form, Request
from fastapi_hotwire import HotwireTemplates, TurboStreamResponse, streams
app = FastAPI()
templates = HotwireTemplates(directory="templates", flashes=False)
todos: list[dict] = []
@app.get("/")
def index(request: Request):
return templates.TemplateResponse(request, "index.html", {"todos": todos})
@app.post("/todos")
def create(request: Request, text: str = Form(...)):
todo = {"id": len(todos) + 1, "text": text}
todos.append(todo)
return templates.render_stream(
request, "index.html", "todo_row",
action="append", target="todos", todo=todo,
)
@app.post("/todos/{todo_id}/delete")
def delete(todo_id: int):
todos[:] = [t for t in todos if t["id"] != todo_id]
return TurboStreamResponse(streams.remove(target=f"todo-{todo_id}"))
A complete runnable version of this example lives in examples/minimal/.
What's in the box
| Module | What it does |
|---|---|
TurboStreamResponse |
A Response subclass with Content-Type: text/vnd.turbo-stream.html. |
streams |
Pure-function builders for <turbo-stream> actions (append, prepend, replace, update, remove, before, after, refresh). |
TurboContext |
A FastAPI dependency that summarizes how the current request relates to Turbo (frame? stream? top-level visit?). |
HotwireTemplates |
A Jinja2Templates wrapper that adds render_block(...) and render_stream(...), plus an automatic flashes context processor. |
flash / get_flashed |
Session-backed flash messages with both a redirect-style and a Hotwire-native turbo-stream flow. |
forms |
A Pydantic ValidationError → turbo-stream renderer for in-place form validation. |
testing |
pytest assertions and request helpers (assert_turbo_stream, parse_streams, assert_turbo_frame, turbo_frame_request, turbo_stream_request). |
TurboStreamResponse
from fastapi_hotwire import TurboStreamResponse, streams
@app.post("/items")
def create():
return TurboStreamResponse([
streams.append(item_html, target="items"),
streams.update(counter_html, target="item-count"),
])
Pass a single string, a list of strings, or None. The class also works as response_class=TurboStreamResponse so OpenAPI documents the media type.
streams
Each builder returns a markupsafe.Markup so it composes safely with Jinja templates:
from fastapi_hotwire import streams
streams.append("<li>...</li>", target="todos")
streams.replace(form_html, target="contact-form")
streams.remove(target="todo-42")
streams.refresh() # Turbo 8 page-refresh
The html argument is interpolated verbatim into the <template> envelope. It must be safe HTML (Jinja autoescaped output is safe). Attribute values (target=, targets=) are HTML-escaped automatically.
TurboContext
from typing import Annotated
from fastapi import Depends
from fastapi_hotwire import TurboContext, turbo_context
@app.post("/items")
async def create(turbo: Annotated[TurboContext, Depends(turbo_context)]):
if turbo.is_frame:
return frame_response(...)
if turbo.accepts_stream:
return stream_response(...)
return full_page_response(...)
Fields: is_frame, frame_id, accepts_stream, is_visit.
HotwireTemplates
from fastapi_hotwire import HotwireTemplates
templates = HotwireTemplates(directory="templates")
@app.get("/items/{id}")
def item_frame(request: Request, id: int):
# Render only the {% block item %} of items.html — useful for
# responding to a <turbo-frame src="..."> request.
return templates.render_block(request, "items.html", "item", item=load(id))
@app.post("/items")
def create(request: Request, text: str = Form(...)):
item = save(text)
# Render the {% block item %} as a turbo-stream that appends to #items.
return templates.render_stream(
request, "items.html", "item",
action="append", target="items", item=item,
)
The flashes context processor is registered automatically; pass flashes=False to opt out.
flash
from fastapi_hotwire import flash, get_flashed
# 1. Classic post-redirect-get flow:
@app.post("/save")
def save(request: Request):
flash(request, "Saved", category="success")
return RedirectResponse("/", status_code=303)
# 2. Hotwire-native: respond with a stream that appends to #flash without redirecting:
@app.post("/save")
def save(request: Request):
return flash.stream(request, "Saved", category="success")
Templates rendered through HotwireTemplates automatically receive the queued flashes list.
A complete runnable example lives in examples/flash/.
forms
from fastapi_hotwire.forms import validation_error_stream
# Render Pydantic validation errors as a turbo-stream that replaces
# only the form's block — no full-page reload, no scroll loss.
try:
Contact.model_validate(form_data)
except ValidationError as exc:
return validation_error_stream(
exc, templates=templates, template="contact.html",
block="form", target="contact-form", request=request,
)
testing
from fastapi_hotwire.testing import (
assert_turbo_stream, parse_streams, assert_turbo_frame,
turbo_frame_request, turbo_stream_request,
)
def test_create_appends_a_row(client):
resp = client.post("/todos", data={"text": "ship"})
assert_turbo_stream(resp)
actions = parse_streams(resp)
assert actions[0].action == "append"
assert actions[0].target == "todos"
Pluggability
fastapi_hotwire.protocols defines the integration seams:
TemplateRenderer— anything implementingrender(name, context) -> str.BlockRenderer— anything implementingrender_block(name, block, context) -> str. The defaultJinja2BlockRendereris one implementation.SessionLike— anyMutableMapping[str, Any]that hangs offrequest.session(StarletteSessionMiddleware,starsessions, an in-memorydict).
You don't need to use the bundled Jinja2 / Starlette code paths to use this library.
Examples
Two full runnable examples live under examples/:
examples/minimal/— A todo list with turbo-stream append + remove. The simplest possible integration.examples/flash/— Session-backed flash messages, with both a PRG and a Hotwire-native flow.
Run either with uv run uvicorn app:app --reload from inside the example directory.
Non-goals
fastapi-hotwire deliberately does not:
- Push to clients via WebSocket / SSE — Hotwire's broadcast /
turbo_stream_frompatterns belong in app code with your message bus of choice. - Bundle a Stimulus JavaScript distribution — load Stimulus the way you load any other JS.
- Inject logging / observability / tracing — those are application concerns, not library concerns.
- Replace
request.url_for(...)with anything more magical.
This list will not grow.
Contributing
See CONTRIBUTING.md. Issues and PRs welcome — please file an issue first for anything bigger than a typo so we can align on scope.
This project follows the Contributor Covenant Code of Conduct.
License
MIT © 2026 Pyre
Metadata
Release files for fastapi-hotwire 0.3.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| fastapi_hotwire-0.3.0.tar.gz | 62.9 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| fastapi_hotwire-0.3.0-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 81.8 kB
Release files / fastapi_hotwire-0.3.0.tar.gz
| Download URL | fastapi_hotwire-0.3.0.tar.gz |
|---|---|
| Size | 62.9 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
90f0b5dd2c5f7598f9975e800f59a581a07c2705b1f81f1126afdb3d972d901a
|
|
BLAKE2b-256 checksum How to use checksums |
88ebf9a5ba4047a04b4c43f300c5ef1cd862bf787d63802c650dc4e8aef82366
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.13
|
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 15, 2026.
Transparency logRelease files / fastapi_hotwire-0.3.0-py3-none-any.whl
| Download URL | fastapi_hotwire-0.3.0-py3-none-any.whl |
|---|---|
| Size | 18.9 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
87b65b7dc7654b63c67f3eded03403b5811a4abbf7b2ad73ddc01a084b153e3f
|
|
BLAKE2b-256 checksum How to use checksums |
cbf9e4b74ecc453751c78c4576320804775ab350477eb594cefd9c4cce0bd13f
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.13
|
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 15, 2026.
Transparency log