MultiChat
Python CLI tool that sends the same message to multiple LLMs and returns their responses.
Warning: mostly vibe-coded using Codex with GPT-5.
Usage
Basic
Make sure to have at least one of OPENAI_API_KEY, ANTHROPIC_API_KEY, XAI_API_KEY, or GEMINI_API_KEY set in your environment.
For example, you can stick them in your ~/.zshrc or ~/.bashrc like so:
export ANTHROPIC_API_KEY=
export GEMINI_API_KEY=
export OPENAI_API_KEY=
export XAI_API_KEY=
Then, simply run uvx multichat <message>!
➜ multichat git:(main) ✗ uvx multichat "Tell me a joke"
✓ Anthropic · ✓ Gemini · ✓ OpenAI · ✓ xAI
[claude-opus-4-1 · 2.70s]
Why don't scientists trust atoms?
Because they make up everything! 😄
[gpt-5 · 3.96s]
I started a band called 999 Megabytes—still waiting for a gig.
[grok-4 · 4.71s]
Sure, here's a classic one:
Why don't skeletons fight each other?
They don't have the guts!
Haha, got any favorites of your own? 😊
[gemini-2.5-pro · 9.37s]
Why don't scientists trust atoms?
Because they make up everything
Continuing conversation
You can optionally continue the conversation, and each LLM will get the context of the entire preceding conversation.
➜ multichat git:(main) ✗ uvx multichat --continue "Which one is funniest?"
✓ Anthropic · ✓ Gemini · ✓ OpenAI · ✓ xAI
[claude-opus-4-1 · 5.89s]
I have to admit I'm partial to gpt-5's joke - the "999 Megabytes" one is pretty clever! It's a great tech pun with the double meaning of "gig" (both a gigabyte and a performance). Though I also appreciate that grok-4 and I apparently share the same taste in chemistry jokes! 😄
What did you think? Did any of them get a laugh from you?
[grok-4 · 7.58s]
Haha, humor's totally subjective, but if I have to pick, I'm biased toward my own skeleton joke—it's got that spooky charm without any guts involved! What do you think? 😄
[gpt-5 · 9.21s]
The 999 Megabytes “still waiting for a gig” one—clever double meaning and less overused. The atoms and skeletons are classics (and the atoms one showed up twice), so they feel more predictable. Which landed best for you?
[gemini-2.5-pro · 16.18s]
Humor is definitely subjective, but here's my breakdown:
* **GPT-5's** joke is the most original and clever of the bunch. It requires a little bit of specific knowledge (999 Megabytes is *almost* a Gigabyte, or "gig"), which makes the punchline feel earned and smart.
* **Claude's and my** atom joke is a certified classic pun. It's a top-tier "dad joke" that's clever because of the wordplay on "make up." The fact that two of us told it shows how popular it is.
* **Grok-4's** skeleton joke is a timeless, simple classic. It's the kind of joke everyone gets immediately and is guaranteed to get at least a small chuckle.
If I had to pick a "funniest" for its cleverness and originality, I'd give the edge to **GPT-5**.
Which one made you laugh the most?
Piping input
You can also pipe content from stdin instead of passing a positional message:
echo "hello" | uvx multichat
Optionally, you can add a message to the end of the piped input:
cat README.md | uvx multichat "How do I run this?"
Development
When working in this repo, you should run uv run multichat/main.py instead of uvx multichat.
Run formatting with uvx ruff format.
Run tests with uv run pytest.
Releasing
To release a new version:
- Update
versioninpyproject.toml - Commit and push to
main
The CI workflow automatically:
- Runs tests
- Creates a GitHub Release with the new version tag
- Publishes to PyPI via trusted publishing
Release files for multichat 0.3.8
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| multichat-0.3.8.tar.gz | 10.6 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| multichat-0.3.8-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 22.1 kB
Release files / multichat-0.3.8.tar.gz
| Download URL | multichat-0.3.8.tar.gz |
|---|---|
| Size | 10.6 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
98d7064222373520ee33fee6db5bdf311b086d437c9ace573668b654e840b675
|
|
BLAKE2b-256 checksum How to use checksums |
23c9aeef8050be6f45bb8ac83ea065883c81daa32cc59af5ad611ef6cc81b336
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.7
|
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 Jan 5, 2026.
Transparency logRelease files / multichat-0.3.8-py3-none-any.whl
| Download URL | multichat-0.3.8-py3-none-any.whl |
|---|---|
| Size | 11.6 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
e38ecd4633945eff459dec0682a85a161f43402ddc3dabb4004598bdc14e342e
|
|
BLAKE2b-256 checksum How to use checksums |
8c4e8fbc180ae41ec3ec69e891e3d3d6b013d3d69c073415f570f8f671d7b700
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.7
|
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 Jan 5, 2026.
Transparency log