Skip to main content

termflow

A streaming markdown renderer and terminal UI toolkit for modern terminals.

PyPI version Python versions License: MIT

termflow renders markdown to ANSI as it arrives, line by line, which makes it a natural fit for LLM output. It also ships the surrounding machinery a terminal-native application needs: smooth typewriter-style output pacing, terminal-wide color theming, and dependency-free interactive menus.

Two runtime dependencies: Pygments and wcwidth. No curses, no prompt_toolkit, no Rich.

Features

  • Streaming rendering — parse and render markdown incrementally, line by line, without waiting for the full document
  • Syntax highlighting — fenced code blocks highlighted via Pygments, with language detection
  • GitHub-flavored tables, ordered/unordered/nested lists, block quotes, and <think> blocks for LLM chain-of-thought
  • Smooth output pacing (termflow.stream) — adaptive-rate buffering that turns bursty token streams into steady typewriter output
  • Terminal theming (termflow.themes) — bundled 16-color palettes applied terminal-wide via OSC escape sequences, with automatic restore on exit
  • Interactive menus (termflow.tui) — a declarative menu builder with search, pagination, multi-select, and live preview panes, built on plain ANSI escape codes
  • OSC 8 hyperlinks and OSC 52 clipboard integration where the terminal supports them
  • Configurable via TOML config file or programmatic API

Installation

pip install termflow-md

Or run the CLI directly:

uvx --from termflow-md tf README.md

CLI

tf README.md                  # render a file
echo "# Hello" | tf           # render stdin
tf -w 100 document.md         # fixed width
tf --style dracula README.md  # color preset
tf --syntax-style nord doc.md # Pygments style for code blocks
tf --list-syntax-styles       # available syntax styles

Run tf --help for the full option list.

Rendering markdown

from termflow import render_markdown

render_markdown("# Hello World")

Streaming, the primary use case:

import sys
from termflow import Parser, Renderer

parser = Parser()
renderer = Renderer(output=sys.stdout, width=80)

for line in markdown_stream:
    renderer.render_all(parser.parse_line(line))

renderer.render_all(parser.finalize())

Custom styling:

from termflow import Renderer, RenderStyle, RenderFeatures

style = RenderStyle.dracula()  # or .nord(), .gruvbox(), .default()
style = RenderStyle(bright="#87ceeb")  # or roll your own

renderer = Renderer(
    width=100,
    style=style,
    features=RenderFeatures(clipboard=True, hyperlinks=True),
)

Smooth streaming output

Token streams arrive in bursts; printing each chunk immediately makes output stutter. termflow.stream buffers incoming text and drains it at an adaptive rate from a background asyncio task: latency stays low when the producer runs hot, and output stays smooth when it trickles.

SmoothWriter is a file-like proxy that sits between a Renderer (or any producer of ANSI text) and the real output stream. Escape sequences are emitted atomically, so styling never tears mid-sequence:

import sys
from termflow import Parser, Renderer
from termflow.stream import SmoothWriter

writer = SmoothWriter(sys.stdout)
writer.start()

renderer = Renderer(output=writer, width=80)
parser = Parser()
async for chunk in model_stream:
    renderer.render_all(parser.parse_line(chunk))

await writer.close()  # waits for the buffer to finish draining
# writer.abort()       # or: stop typing NOW and drop the backlog

StreamSmoother does the same for plain text via an emit callback, and both accept an is_paused hook to hold output while something else owns the terminal.

Terminal theming

termflow.themes recolors the whole terminal window — background, foreground, and the 16 ANSI palette slots — using xterm OSC sequences supported by iTerm2, Terminal.app, kitty, Alacritty, VS Code, GNOME Terminal, and Windows Terminal. Unsupported terminals ignore them silently. An atexit handler restores the terminal on process exit.

from termflow.themes import PALETTES, apply_palette, reset_palette

apply_palette(PALETTES["catppuccin_mocha"])
reset_palette()  # back to the terminal's own colors

Bundled palettes: Catppuccin Mocha/Latte, Tokyo Night, Solarized Light, GitHub Light, Rose Pine Dawn, and a set of originals (ocean, forest, sunset, vaporwave, green_screen, deep_black, purple_puppy, bubblegum_pink).

Each palette bridges to the markdown renderer, so themed output matches the terminal chrome:

from termflow import Renderer
from termflow.themes import get_palette

palette = get_palette("tokyo_night")
renderer = Renderer(style=palette.to_render_style())

Interactive menus

termflow.tui provides a menu component built on raw ANSI escape codes: alternate screen, arrow-key navigation, incremental search, pagination, multi-select, and a live preview pane. Every I/O surface (key source, output stream, terminal size) is injectable, so menus are testable without a tty.

from termflow.tui import MenuBuilder, MenuItem

result = (
    MenuBuilder("Pick a model")
    .items(
        [
            MenuItem("gpt-5", description="fast and smart"),
            MenuItem("claude", description="thoughtful"),
            MenuItem("qwen", description="local"),
        ]
    )
    .searchable()
    .page_size(10)
    .preview(lambda item: f"Details for {item.label}")
    .run()
)

if not result.cancelled:
    print(result.item.value)

Multi-select returns result.items; on_highlight fires on every cursor move (useful for live theme previews); disabled items render dim and are skipped by navigation.

Configuration

Create ~/.config/termflow/config.toml (or point TERMFLOW_CONFIG at a path):

width = 0            # 0 = auto-detect
max_width = 120
syntax_style = "monokai"

[style]
bright = "#87ceeb"   # main accent (H1/H2)
head = "#98fb98"     # H3
symbol = "#dda0dd"   # bullets, borders
link = "#87cefa"
error = "#ff6b6b"

[features]
clipboard = true     # OSC 52 clipboard for code blocks
hyperlinks = true    # OSC 8 clickable links
pretty_pad = true    # unicode borders on code blocks

See examples/config.toml for the full set of options.

Origin

termflow began as a Python port of streamdown-rs, a streaming markdown renderer written in Rust, and has since grown into a broader terminal UI toolkit.

Contributing

git clone https://github.com/mpfaffenberger/termflow.git
cd termflow
pip install -e ".[dev]"

pytest tests/ -v
ruff check .
ruff format .

Pull requests are welcome.

License

MIT. See LICENSE.

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

termflow_md-0.3.0.tar.gz (146.7 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

termflow_md-0.3.0-py3-none-any.whl (84.4 kB view details)

Uploaded Python 3

File details

Details for the file termflow_md-0.3.0.tar.gz.

File metadata

  • Download URL: termflow_md-0.3.0.tar.gz
  • Upload date:
  • Size: 146.7 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/7.0.0 CPython/3.12.3

File hashes

Hashes for termflow_md-0.3.0.tar.gz
Algorithm Hash digest
SHA256 a17d290a3f77449035847e265c0ac054af04bb2a85cd93bc725085ee79332eb1
MD5 6a905ff26104736a0324ff2f488cdfe8
BLAKE2b-256 7d89f8e8bd25846a6d34a3d230302dac7a91df32c9f59a308698a4a9461d6481

See more details on using hashes here.

File details

Details for the file termflow_md-0.3.0-py3-none-any.whl.

File metadata

  • Download URL: termflow_md-0.3.0-py3-none-any.whl
  • Upload date:
  • Size: 84.4 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/7.0.0 CPython/3.12.3

File hashes

Hashes for termflow_md-0.3.0-py3-none-any.whl
Algorithm Hash digest
SHA256 2c6721248e7c4a19b0bbf474e7bbb32c7c69e93e2e1084e30f6bf7f992a9c8a2
MD5 7def899a93450d6e423101a092cacc4a
BLAKE2b-256 69ec723e47f8a5c6f14a3cb8714b78d93dc80e1eb0be3304f0201d616eb3826c

See more details on using hashes here.

Release history Release notifications | RSS feed

0.9.1

2 files

0.9.0

2 files

0.8.0

2 files

0.7.0

2 files

0.6.0

2 files

0.5.1

2 files

0.5.0

2 files

0.4.0

2 files

This release

0.3.0 This release

2 files

0.2.6

2 files

0.2.5

2 files

0.2.4

2 files

0.2.3

2 files

0.2.2

2 files

0.2.1

2 files

0.2.0

2 files

0.1.11

2 files

0.1.10

2 files

0.1.9

2 files

0.1.8

2 files

0.1.7

2 files

0.1.6

2 files

0.1.0

2 files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page