Getting Started

June 20, 2026 · View on GitHub

This guide walks through the fastest way to see CANarchy produce familiar candump-style output.

Install

CANarchy targets Python 3.12 or newer. Pick the install path that matches your use case.

Windows 11 quickstart

Use Windows Terminal with PowerShell 7 when possible. The commands below also work in Windows PowerShell 5.1 unless noted. cmd.exe is usable for CANarchy commands, but PowerShell is clearer for environment variables and JSON output.

  1. Install Python 3.12 or newer from python.org or with winget:

    winget install Python.Python.3.12
py -3.12 --version

  2. Install pipx, then make sure its script directory is on PATH:

    py -3.12 -m pip install --user pipx
py -3.12 -m pipx ensurepath

    Open a new terminal after ensurepath if pipx is not found immediately.

  3. Install CANarchy from PyPI and run the offline health check:

    pipx install canarchy
canarchy --version
canarchy doctor --text

    canarchy doctor should print a normal health-check table on Windows. Warnings about missing hardware-specific drivers or an unpopulated DBC cache are expected until you configure those workflows.

  4. Run the deterministic no-hardware smoke test:

    $env:CANARCHY_TRANSPORT_BACKEND = "scaffold"
canarchy capture can0 --jsonl

    In cmd.exe, use this environment-variable syntax instead:

    set CANARCHY_TRANSPORT_BACKEND=scaffold
canarchy capture can0 --jsonl

  5. For live CAN hardware, install the vendor driver first, then select the matching python-can interface type. Common Windows paths are:

    Adapter familyDriver pointerCANarchy interface typeExample channel
    Vector VN/VXVector Driver Setupvector0
    PEAK PCAN-USB/PCIePEAK-System driverspcanPCAN_USBBUS1

    Example PCAN session configuration:

    $env:CANARCHY_PYTHON_CAN_INTERFACE = "pcan"
canarchy config show
canarchy capture PCAN_USBBUS1 --candump

    Example Vector session configuration:

    $env:CANARCHY_PYTHON_CAN_INTERFACE = "vector"
canarchy config show
canarchy capture 0 --candump

Known Windows limitations:

  • SocketCAN and vcan are Linux-only; use vendor hardware backends, udp_multicast, or scaffold on Windows.
  • canarchy completion currently emits bash, zsh, and fish scripts, not PowerShell completion.
  • The virtual python-can interface is same-process only. Use udp_multicast for no-hardware tests that need separate terminals.
  • Live capture/transmit depends on vendor drivers, adapter firmware, and channel names exposed by the driver package.
pipx install canarchy            # isolated, on PATH everywhere
# or
pip install --user canarchy      # fallback if pipx is unavailable

Confirm the install:

canarchy --version
canarchy doctor --text

canarchy doctor runs offline health checks (Python, python-can, transport backend, configured python-can vendor interface imports, caches, config file, MCP server) — every check green means the software environment is ready.

From source (development)

CANarchy uses uv for environment, dependency, and packaging workflows.

uv sync                          # create .venv, install everything

Then pick one of:

  • Install as a global tool (no uv run prefix needed afterwards):

    uv tool install --editable .

    Source edits take effect without reinstalling. Uninstall later with uv tool uninstall canarchy.

  • Activate the project virtualenv for the current shell session:

    source .venv/bin/activate

    Pair with direnv to activate automatically on cd.

The rest of this guide assumes canarchy is available directly. Prefix commands with uv run if you have not done either step above.

Live Capture

CANarchy defaults to the python-can backend with SocketCAN, so on Linux with a configured CAN interface you can start capturing immediately — no config file needed:

canarchy capture can0 --candump

This command stays open and prints frames in candump format until you interrupt it with Ctrl+C.

Use --json or --jsonl for machine-readable output:

canarchy capture can0 --jsonl

Verify your active configuration

canarchy config show

This prints each setting along with whether it came from a config file, an environment variable, or the built-in default.

A fully commented sample configuration file is available at docs/examples/config.toml. Copy it to ~/.canarchy/config.toml and uncomment the keys you want to set.

Check the local environment

canarchy doctor runs a battery of fast, offline health checks (Python version, python-can import, transport backend, configured python-can vendor interface imports, cache writability, opendbc cache, MCP server, config file). Run it before reporting bugs:

canarchy doctor --text

Each check shows [OK], [WARN], or [FAIL] and includes a copy-pasteable remediation hint for non-ok results.

For hardware backends such as PCAN, Vector, and Kvaser, doctor checks software dependencies only. Use the Backends & Interfaces page and the cookbook verification recipes for live adapter validation.

Install shell completion

Tab-completion scripts for bash, zsh, and fish are emitted by the completion subcommand. Pick the snippet for your shell:

# bash
eval "$(canarchy completion bash)"            # one-off in current shell
canarchy completion bash > ~/.bash_completion.d/canarchy   # persistent

# zsh
eval "$(canarchy completion zsh)"             # one-off in current shell
canarchy completion zsh > ~/.zsh/completions/_canarchy && compinit

# fish
canarchy completion fish | source             # one-off in current shell
canarchy completion fish > ~/.config/fish/completions/canarchy.fish

The scripts complete the top-level subcommands and the most common flags (--json, --jsonl, --text, --file, --dbc, --max-frames, --seconds, --offset, --ack-active, --log-level, --quiet).

Software Loopback (No Hardware)

For a local send-and-receive loop across two terminals without real hardware, use the udp_multicast interface. The virtual interface is in-process only and will not deliver frames between separate terminal sessions.

On macOS you may need to add a multicast route first:

sudo route add -net 239.0.0.0/8 -interface lo0

Set udp_multicast in ~/.canarchy/config.toml:

[transport]
backend = "python-can"
interface = "udp_multicast"
default_interface = "239.0.0.1"

Or export for the current session only:

export CANARCHY_PYTHON_CAN_INTERFACE=udp_multicast

In one terminal, start a live candump capture:

canarchy capture 239.0.0.1 --candump

In another terminal, send a frame:

canarchy send 239.0.0.1 0x123 11223344 --json

If default_interface is configured, you can omit the repeated channel:

canarchy send 0x123 11223344 --json

The capture terminal should print a line like:

(1713369600.000000) 239.0.0.1 123#11223344

Offline / scaffold mode

To run CANarchy without any CAN interface (for demos, testing, or CI), force the scaffold backend:

export CANARCHY_TRANSPORT_BACKEND=scaffold
canarchy capture can0 --json   # returns pre-recorded fixture frames

Or set it persistently in ~/.canarchy/config.toml:

[transport]
backend = "scaffold"

Read an Existing Candump File

CANarchy can operate on standard timestamped candump log files.

Summarize a capture:

canarchy stats --file tests/fixtures/sample.candump --json

Filter for a specific arbitration ID:

canarchy filter 'id==0x18FEEE31' --file tests/fixtures/sample.candump --json

Replay a capture with deterministic timing:

canarchy replay --file tests/fixtures/sample.candump --rate 1.0 --json

Analyze CANdid Dataset Files

The CANdid dataset (VehicleSec 2025) provides candump-format CAN logs from 10 passenger vehicles.

Replay the default CANdid catalog stream directly from the remote provider without downloading the full file first:

canarchy datasets replay catalog:candid --dry-run --json
canarchy datasets replay catalog:candid --rate 1.0
canarchy datasets replay catalog:candid --format jsonl --rate 10 --max-frames 1000
canarchy datasets replay catalog:candid --list-files --json
canarchy datasets replay catalog:candid --file 2_indicator_CAN.log --rate 1000 --max-frames 10

Remote replay writes candump or JSONL records to stdout unless --json is requested. Use --list-files --json to inspect the embedded CANdid replay manifest, then pass --file <id-or-name> to select a specific replay file.

!!! tip "Acquire quickly, analyze bounded" datasets replay is real-time by default (--rate 1.0), and datasets fetch records provenance only — it does not download frames. To pull a fast, bounded slice raise --rate and cap with --max-frames/--max-seconds (note: datasets replay uses --max-seconds, not --seconds). For per-frame decoders on large files, run capture-info first and pass its suggested_max_frames/suggested_seconds to the command's --max-frames/--seconds. See Working with large captures and Acquiring dataset data.

You can also pipe remote candump replay directly into stdin-aware analysis commands without creating a temporary file:

canarchy datasets replay catalog:candid --rate 1000 --max-frames 100 \
  | canarchy stats --file - --json
canarchy datasets replay catalog:candid --rate 1000 --max-frames 100 \
  | canarchy capture-info --file - --json

After downloading specific *_CAN.log files, use them directly with file-backed analysis commands:

Summarize a CANdid capture:

canarchy stats --file 2_driving_CAN.log --json
canarchy capture-info --file 2_driving_CAN.log --json

Filter for specific arbitration IDs:

canarchy filter 'id==0x123' --file 2_steering_CAN.log --json

Run reverse-engineering helpers:

canarchy re entropy --file 2_driving_CAN.log
canarchy re counters --file 2_driving_CAN.log

The catalog entry includes metadata about annotations, GPS traces, and video artifacts:

canarchy datasets inspect catalog:candid --json

Discover More Public CAN Dataset Sources

CANarchy also documents curated dataset indexes, such as the PIVOT Auto dataset page, that point to multiple external CAN, CAN-FD, J1939, intrusion-detection, and automotive research datasets.

Search for the PIVOT Auto catalog entry:

canarchy datasets search pivot
canarchy datasets inspect catalog:pivot-auto-datasets --json

The PIVOT entry is an index, not a directly replayable dataset. Follow the linked source pages for downloads, licenses, and file formats, then use the relevant CANarchy analysis workflow for the files you obtain.

Discover and Use DBC Files

CANarchy can work with local DBC files or provider-backed refs.

Search the optional opendbc catalog:

canarchy dbc cache refresh --provider opendbc
canarchy dbc search toyota --provider opendbc --limit 5 --json

Decode using a provider ref instead of a local file path:

canarchy decode --file tests/fixtures/sample.candump --dbc opendbc:toyota_tnga_k_pt_generated --json

Structured output for decode, encode, and dbc inspect includes a data.dbc_source object so you can see which local or provider-backed DBC resolution was used.

Supported Candump Forms

Current supported candump text forms include:

  • classic frames: 123#11223344
  • remote frames: 123#R
  • CAN FD frames: 123##31122334455667788
  • error frames using a CAN error-flagged identifier such as 20000080#0000000000000000

For the full current command surface, see the Command Spec.

For a complete list of supported python-can interface types and per-adapter setup instructions, see Backends & Interfaces.