1**NOTE:** This is a personal fork of
2[Crush](https://github.com/charmbracelet/crush) where I'm experimenting with
3things. You'll also find my patches merged here before they're merged into
4upstream. I've also unconditionally disabled metrics and yote the CLA; devs
5shouldn't assign their copyrights to maintainers and hand them the rug they'll
6yank from under you later.
7
8Install my fork either…
9
10- Using [bin](https://github.com/marcosnils/bin) (highly recommended
11 because it's one tool to manage and update myriad CLI tools
12 distributed as statically-linked binaries, like _nasin pali_)
13 ```bash
14 bin install goinstall://git.secluded.site/crush@v0.29.0-fork.2
15 ```
16- Or using the [go toolchain](https://go.dev/dl) (requires tracking updates
17 manually)
18 ```bash
19 go install git.secluded.site/crush@v0.29.0-fork.2
20 ```
21
22I pull upstream changes into the `upstream` branch here, break my
23branches off that, then merge them into my `dev` branch, which gets
24rebased on `upstream` fairly often. `dev` history will get rewritten
25frequently.
26
27Contributions are welcome (see following section) and I'll try to
28maintain anything I accept here.
29
30## Contributions
31
32Patch requests are in [amolith/llm-projects] on [pr.pico.sh]. You don't
33need a new account to contribute, you don't need to fork this repo, you
34don't need to fiddle with `git send-email`, you don't need to faff with
35your email client to get `git request-pull` working...
36
37You just need:
38
39- Git
40- SSH
41- An SSH key
42
43```sh
44# Clone this repo, make your changes, and commit them
45# Create a new patch request with
46git format-patch origin/main --stdout | ssh pr.pico.sh pr create amolith/llm-projects
47# After potential feedback, submit a revision to an existing patch request with
48git format-patch origin/main --stdout | ssh pr.pico.sh pr add {prID}
49# List patch requests
50ssh pr.pico.sh pr ls amolith/llm-projects
51```
52
53See "How do Patch Requests work?" on [pr.pico.sh]'s home page for a more
54complete example workflow.
55
56[amolith/llm-projects]: https://pr.pico.sh/r/amolith/llm-projects
57[pr.pico.sh]: https://pr.pico.sh
58
59---
60
61_Everything from here on is from the original Crush README._
62
63---
64
65# Crush
66
67<p align="center">
68 <a href="https://stuff.charm.sh/crush/charm-crush.png"><img width="450" alt="Charm Crush Logo" src="https://github.com/user-attachments/assets/adc1a6f4-b284-4603-836c-59038caa2e8b" /></a><br />
69 <a href="https://github.com/charmbracelet/crush/releases"><img src="https://img.shields.io/github/release/charmbracelet/crush" alt="Latest Release"></a>
70 <a href="https://github.com/charmbracelet/crush/actions"><img src="https://github.com/charmbracelet/crush/actions/workflows/build.yml/badge.svg" alt="Build Status"></a>
71</p>
72
73<p align="center">Your new coding bestie, now available in your favourite terminal.<br />Your tools, your code, and your workflows, wired into your LLM of choice.</p>
74<p align="center">你的新编程伙伴，现在就在你最爱的终端中。<br />你的工具、代码和工作流，都与您选择的 LLM 模型紧密相连。</p>
75
76<p align="center"><img width="800" alt="Crush Demo" src="https://github.com/user-attachments/assets/58280caf-851b-470a-b6f7-d5c4ea8a1968" /></p>
77
78## Features
79
80- **Multi-Model:** choose from a wide range of LLMs or add your own via OpenAI- or Anthropic-compatible APIs
81- **Flexible:** switch LLMs mid-session while preserving context
82- **Session-Based:** maintain multiple work sessions and contexts per project
83- **LSP-Enhanced:** Crush uses LSPs for additional context, just like you do
84- **Extensible:** add capabilities via MCPs (`http`, `stdio`, and `sse`)
85- **Works Everywhere:** first-class support in every terminal on macOS, Linux, Windows (PowerShell and WSL), FreeBSD, OpenBSD, and NetBSD
86
87## Installation
88
89Use a package manager:
90
91```bash
92# Homebrew
93brew install charmbracelet/tap/crush
94
95# NPM
96npm install -g @charmland/crush
97
98# Arch Linux (btw)
99yay -S crush-bin
100
101# Nix
102nix run github:numtide/nix-ai-tools#crush
103```
104
105Windows users:
106
107```bash
108# Winget
109winget install charmbracelet.crush
110
111# Scoop
112scoop bucket add charm https://github.com/charmbracelet/scoop-bucket.git
113scoop install crush
114```
115
116<details>
117<summary><strong>Nix (NUR)</strong></summary>
118
119Crush is available via [NUR](https://github.com/nix-community/NUR) in `nur.repos.charmbracelet.crush`.
120
121You can also try out Crush via `nix-shell`:
122
123```bash
124# Add the NUR channel.
125nix-channel --add https://github.com/nix-community/NUR/archive/main.tar.gz nur
126nix-channel --update
127
128# Get Crush in a Nix shell.
129nix-shell -p '(import <nur> { pkgs = import <nixpkgs> {}; }).repos.charmbracelet.crush'
130```
131
132### NixOS & Home Manager Module Usage via NUR
133
134Crush provides NixOS and Home Manager modules via NUR.
135You can use these modules directly in your flake by importing them from NUR. Since it auto detects whether its a home manager or nixos context you can use the import the exact same way :)
136
137```nix
138{
139 inputs = {
140 nixpkgs.url = "github:NixOS/nixpkgs/nixos-unstable";
141 nur.url = "github:nix-community/NUR";
142 };
143
144 outputs = { self, nixpkgs, nur, ... }: {
145 nixosConfigurations.your-hostname = nixpkgs.lib.nixosSystem {
146 system = "x86_64-linux";
147 modules = [
148 nur.modules.nixos.default
149 nur.repos.charmbracelet.modules.crush
150 {
151 programs.crush = {
152 enable = true;
153 settings = {
154 providers = {
155 openai = {
156 id = "openai";
157 name = "OpenAI";
158 base_url = "https://api.openai.com/v1";
159 type = "openai";
160 api_key = "sk-fake123456789abcdef...";
161 models = [
162 {
163 id = "gpt-4";
164 name = "GPT-4";
165 }
166 ];
167 };
168 };
169 lsp = {
170 go = { command = "gopls"; enabled = true; };
171 nix = { command = "nil"; enabled = true; };
172 };
173 options = {
174 context_paths = [ "/etc/nixos/configuration.nix" ];
175 tui = { compact_mode = true; };
176 debug = false;
177 };
178 };
179 };
180 }
181 ];
182 };
183 };
184}
185```
186
187</details>
188
189<details>
190<summary><strong>Debian/Ubuntu</strong></summary>
191
192```bash
193sudo mkdir -p /etc/apt/keyrings
194curl -fsSL https://repo.charm.sh/apt/gpg.key | sudo gpg --dearmor -o /etc/apt/keyrings/charm.gpg
195echo "deb [signed-by=/etc/apt/keyrings/charm.gpg] https://repo.charm.sh/apt/ * *" | sudo tee /etc/apt/sources.list.d/charm.list
196sudo apt update && sudo apt install crush
197```
198
199</details>
200
201<details>
202<summary><strong>Fedora/RHEL</strong></summary>
203
204```bash
205echo '[charm]
206name=Charm
207baseurl=https://repo.charm.sh/yum/
208enabled=1
209gpgcheck=1
210gpgkey=https://repo.charm.sh/yum/gpg.key' | sudo tee /etc/yum.repos.d/charm.repo
211sudo yum install crush
212```
213
214</details>
215
216Or, download it:
217
218- [Packages][releases] are available in Debian and RPM formats
219- [Binaries][releases] are available for Linux, macOS, Windows, FreeBSD, OpenBSD, and NetBSD
220
221[releases]: https://github.com/charmbracelet/crush/releases
222
223Or just install it with Go:
224
225```
226go install github.com/charmbracelet/crush@latest
227```
228
229> [!WARNING]
230> Productivity may increase when using Crush and you may find yourself nerd
231> sniped when first using the application. If the symptoms persist, join the
232> [Discord][discord] and nerd snipe the rest of us.
233
234## Getting Started
235
236The quickest way to get started is to grab an API key for your preferred
237provider such as Anthropic, OpenAI, Groq, or OpenRouter and just start
238Crush. You'll be prompted to enter your API key.
239
240That said, you can also set environment variables for preferred providers.
241
242| Environment Variable | Provider |
243| --------------------------- | -------------------------------------------------- |
244| `ANTHROPIC_API_KEY` | Anthropic |
245| `OPENAI_API_KEY` | OpenAI |
246| `OPENROUTER_API_KEY` | OpenRouter |
247| `GEMINI_API_KEY` | Google Gemini |
248| `CEREBRAS_API_KEY` | Cerebras |
249| `HF_TOKEN` | Huggingface Inference |
250| `VERTEXAI_PROJECT` | Google Cloud VertexAI (Gemini) |
251| `VERTEXAI_LOCATION` | Google Cloud VertexAI (Gemini) |
252| `GROQ_API_KEY` | Groq |
253| `AWS_ACCESS_KEY_ID` | Amazon Bedrock (Claude) |
254| `AWS_SECRET_ACCESS_KEY` | Amazon Bedrock (Claude) |
255| `AWS_REGION` | Amazon Bedrock (Claude) |
256| `AWS_PROFILE` | Amazon Bedrock (Custom Profile) |
257| `AWS_BEARER_TOKEN_BEDROCK` | Amazon Bedrock |
258| `AZURE_OPENAI_API_ENDPOINT` | Azure OpenAI models |
259| `AZURE_OPENAI_API_KEY` | Azure OpenAI models (optional when using Entra ID) |
260| `AZURE_OPENAI_API_VERSION` | Azure OpenAI models |
261
262### By the Way
263
264Is there a provider you’d like to see in Crush? Is there an existing model that needs an update?
265
266Crush’s default model listing is managed in [Catwalk](https://github.com/charmbracelet/catwalk), a community-supported, open source repository of Crush-compatible models, and you’re welcome to contribute.
267
268<a href="https://github.com/charmbracelet/catwalk"><img width="174" height="174" alt="Catwalk Badge" src="https://github.com/user-attachments/assets/95b49515-fe82-4409-b10d-5beb0873787d" /></a>
269
270## Configuration
271
272Crush runs great with no configuration. That said, if you do need or want to
273customize Crush, configuration can be added either local to the project itself,
274or globally, with the following priority:
275
2761. `.crush.json`
2772. `crush.json`
2783. `$HOME/.config/crush/crush.json`
279
280Configuration itself is stored as a JSON object:
281
282```json
283{
284 "this-setting": { "this": "that" },
285 "that-setting": ["ceci", "cela"]
286}
287```
288
289As an additional note, Crush also stores ephemeral data, such as application state, in one additional location:
290
291```bash
292# Unix
293$HOME/.local/share/crush/crush.json
294
295# Windows
296%LOCALAPPDATA%\crush\crush.json
297```
298
299> [!TIP]
300> You can override the user and data config locations by setting:
301> * `CRUSH_GLOBAL_CONFIG`
302> * `CRUSH_GLOBAL_DATA`
303
304### LSPs
305
306Crush can use LSPs for additional context to help inform its decisions, just
307like you would. LSPs can be added manually like so:
308
309```json
310{
311 "$schema": "https://charm.land/crush.json",
312 "lsp": {
313 "go": {
314 "command": "gopls",
315 "env": {
316 "GOTOOLCHAIN": "go1.24.5"
317 }
318 },
319 "typescript": {
320 "command": "typescript-language-server",
321 "args": ["--stdio"]
322 },
323 "nix": {
324 "command": "nil"
325 }
326 }
327}
328```
329
330### MCPs
331
332Crush also supports Model Context Protocol (MCP) servers through three
333transport types: `stdio` for command-line servers, `http` for HTTP endpoints,
334and `sse` for Server-Sent Events. Environment variable expansion is supported
335using `$(echo $VAR)` syntax.
336
337```json
338{
339 "$schema": "https://charm.land/crush.json",
340 "mcp": {
341 "filesystem": {
342 "type": "stdio",
343 "command": "node",
344 "args": ["/path/to/mcp-server.js"],
345 "timeout": 120,
346 "disabled": false,
347 "disabled_tools": ["some-tool-name"],
348 "env": {
349 "NODE_ENV": "production"
350 }
351 },
352 "github": {
353 "type": "http",
354 "url": "https://api.githubcopilot.com/mcp/",
355 "timeout": 120,
356 "disabled": false,
357 "disabled_tools": ["create_issue", "create_pull_request"],
358 "headers": {
359 "Authorization": "Bearer $GH_PAT"
360 }
361 },
362 "streaming-service": {
363 "type": "sse",
364 "url": "https://example.com/mcp/sse",
365 "timeout": 120,
366 "disabled": false,
367 "headers": {
368 "API-Key": "$(echo $API_KEY)"
369 }
370 }
371 }
372}
373```
374
375### Memory
376
377Crush automatically includes two memory files for cross-project instructions.
378
379- `~/.config/crush/CRUSH.md`: Crush-specific rules that would confuse other
380 agentic coding tools. If you only use Crush, this is the only one you need to
381 edit.
382- `~/.config/AGENTS.md`: generic instructions that other coding tools might
383 read. Avoid referring to Crush-specific tools or workflows here.
384
385You can customize these paths using the `memory_paths` option in your
386configuration:
387
388```json
389{
390 "$schema": "https://charm.land/crush.json",
391 "options": {
392 "memory_paths": [
393 "/path/to/custom/memory/file.md",
394 "/path/to/folder/of/files/" // recursively load all .md files in folder
395 ]
396 }
397}
398```
399
400### Ignoring Files
401
402Crush respects `.gitignore` files by default, but you can also create a
403`.crushignore` file to specify additional files and directories that Crush
404should ignore. This is useful for excluding files that you want in version
405control but don't want Crush to consider when providing context.
406
407The `.crushignore` file uses the same syntax as `.gitignore` and can be placed
408in the root of your project or in subdirectories.
409
410### Allowing Tools
411
412By default, Crush will ask you for permission before running tool calls. If
413you'd like, you can allow tools to be executed without prompting you for
414permissions. Use this with care.
415
416```json
417{
418 "$schema": "https://charm.land/crush.json",
419 "permissions": {
420 "allowed_tools": [
421 "view",
422 "ls",
423 "grep",
424 "edit",
425 "mcp_context7_get-library-doc"
426 ]
427 }
428}
429```
430
431You can also skip all permission prompts entirely by running Crush with the
432`--yolo` flag. Be very, very careful with this feature.
433
434### Disabling Built-In Tools
435
436If you'd like to prevent Crush from using certain built-in tools entirely, you
437can disable them via the `options.disabled_tools` list. Disabled tools are
438completely hidden from the agent.
439
440```json
441{
442 "$schema": "https://charm.land/crush.json",
443 "options": {
444 "disabled_tools": [
445 "bash",
446 "sourcegraph"
447 ]
448 }
449}
450```
451
452To disable tools from MCP servers, see the [MCP config section](#mcps).
453
454### Desktop notifications
455
456Crush sends desktop notifications when a tool call requires permission and when
457the agent finishes its turn. They're only sent when the terminal window isn't
458focused _and_ your terminal supports reporting the focus state.
459
460```json
461{
462 "$schema": "https://charm.land/crush.json",
463 "options": {
464 "disable_notifications": true // default
465 }
466}
467```
468
469To disable desktop notifications, set `disable_notifications` to `true` in your
470configuration. On macOS, notifications currently lack icons due to platform
471limitations.
472
473### Initialization
474
475When you initialize a project, Crush analyzes your codebase and creates
476a context file that helps it work more effectively in future sessions.
477By default, this file is named `AGENTS.md`, but you can customize the
478name and location with the `initialize_as` option:
479
480```json
481{
482 "$schema": "https://charm.land/crush.json",
483 "options": {
484 "initialize_as": "AGENTS.md"
485 }
486}
487```
488
489This is useful if you prefer a different naming convention or want to
490place the file in a specific directory (e.g., `CRUSH.md` or
491`docs/LLMs.md`). Crush will fill the file with project-specific context
492like build commands, code patterns, and conventions it discovered during
493initialization.
494
495### Attribution Settings
496
497By default, Crush adds attribution information to Git commits and pull requests
498it creates. You can customize this behavior with the `attribution` option:
499
500```json
501{
502 "$schema": "https://charm.land/crush.json",
503 "options": {
504 "attribution": {
505 "trailer_style": "co-authored-by",
506 "generated_with": true
507 }
508 }
509}
510```
511
512- `trailer_style`: Controls the attribution trailer added to commit messages
513 (default: `assisted-by`)
514 - `assisted-by`: Adds `Assisted-by: [Model Name] via Crush <crush@charm.land>`
515 (includes the model name)
516 - `co-authored-by`: Adds `Co-Authored-By: Crush <crush@charm.land>`
517 - `none`: No attribution trailer
518- `generated_with`: When true (default), adds `💘 Generated with Crush` line to
519 commit messages and PR descriptions
520
521### Custom Providers
522
523Crush supports custom provider configurations for both OpenAI-compatible and
524Anthropic-compatible APIs.
525
526> [!NOTE]
527> Note that we support two "types" for OpenAI. Make sure to choose the right one
528> to ensure the best experience!
529> * `openai` should be used when proxying or routing requests through OpenAI.
530> * `openai-compat` should be used when using non-OpenAI providers that have OpenAI-compatible APIs.
531
532#### OpenAI-Compatible APIs
533
534Here’s an example configuration for Deepseek, which uses an OpenAI-compatible
535API. Don't forget to set `DEEPSEEK_API_KEY` in your environment.
536
537```json
538{
539 "$schema": "https://charm.land/crush.json",
540 "providers": {
541 "deepseek": {
542 "type": "openai-compat",
543 "base_url": "https://api.deepseek.com/v1",
544 "api_key": "$DEEPSEEK_API_KEY",
545 "models": [
546 {
547 "id": "deepseek-chat",
548 "name": "Deepseek V3",
549 "cost_per_1m_in": 0.27,
550 "cost_per_1m_out": 1.1,
551 "cost_per_1m_in_cached": 0.07,
552 "cost_per_1m_out_cached": 1.1,
553 "context_window": 64000,
554 "default_max_tokens": 5000
555 }
556 ]
557 }
558 }
559}
560```
561
562#### Anthropic-Compatible APIs
563
564Custom Anthropic-compatible providers follow this format:
565
566```json
567{
568 "$schema": "https://charm.land/crush.json",
569 "providers": {
570 "custom-anthropic": {
571 "type": "anthropic",
572 "base_url": "https://api.anthropic.com/v1",
573 "api_key": "$ANTHROPIC_API_KEY",
574 "extra_headers": {
575 "anthropic-version": "2023-06-01"
576 },
577 "models": [
578 {
579 "id": "claude-sonnet-4-20250514",
580 "name": "Claude Sonnet 4",
581 "cost_per_1m_in": 3,
582 "cost_per_1m_out": 15,
583 "cost_per_1m_in_cached": 3.75,
584 "cost_per_1m_out_cached": 0.3,
585 "context_window": 200000,
586 "default_max_tokens": 50000,
587 "can_reason": true,
588 "supports_attachments": true
589 }
590 ]
591 }
592 }
593}
594```
595
596### Amazon Bedrock
597
598Crush currently supports running Anthropic models through Bedrock, with caching disabled.
599
600- A Bedrock provider will appear once you have AWS configured, i.e. `aws configure`
601- Crush also expects the `AWS_REGION` or `AWS_DEFAULT_REGION` to be set
602- To use a specific AWS profile set `AWS_PROFILE` in your environment, i.e. `AWS_PROFILE=myprofile crush`
603- Alternatively to `aws configure`, you can also just set `AWS_BEARER_TOKEN_BEDROCK`
604
605### Vertex AI Platform
606
607Vertex AI will appear in the list of available providers when `VERTEXAI_PROJECT` and `VERTEXAI_LOCATION` are set. You will also need to be authenticated:
608
609```bash
610gcloud auth application-default login
611```
612
613To add specific models to the configuration, configure as such:
614
615```json
616{
617 "$schema": "https://charm.land/crush.json",
618 "providers": {
619 "vertexai": {
620 "models": [
621 {
622 "id": "claude-sonnet-4@20250514",
623 "name": "VertexAI Sonnet 4",
624 "cost_per_1m_in": 3,
625 "cost_per_1m_out": 15,
626 "cost_per_1m_in_cached": 3.75,
627 "cost_per_1m_out_cached": 0.3,
628 "context_window": 200000,
629 "default_max_tokens": 50000,
630 "can_reason": true,
631 "supports_attachments": true
632 }
633 ]
634 }
635 }
636}
637```
638
639### Local Models
640
641Local models can also be configured via OpenAI-compatible API. Here are two common examples:
642
643#### Ollama
644
645```json
646{
647 "providers": {
648 "ollama": {
649 "name": "Ollama",
650 "base_url": "http://localhost:11434/v1/",
651 "type": "openai-compat",
652 "models": [
653 {
654 "name": "Qwen 3 30B",
655 "id": "qwen3:30b",
656 "context_window": 256000,
657 "default_max_tokens": 20000
658 }
659 ]
660 }
661 }
662}
663```
664
665#### LM Studio
666
667```json
668{
669 "providers": {
670 "lmstudio": {
671 "name": "LM Studio",
672 "base_url": "http://localhost:1234/v1/",
673 "type": "openai-compat",
674 "models": [
675 {
676 "name": "Qwen 3 30B",
677 "id": "qwen/qwen3-30b-a3b-2507",
678 "context_window": 256000,
679 "default_max_tokens": 20000
680 }
681 ]
682 }
683 }
684}
685```
686
687## Logging
688
689Sometimes you need to look at logs. Luckily, Crush logs all sorts of
690stuff. Logs are stored in `./.crush/logs/crush.log` relative to the project.
691
692The CLI also contains some helper commands to make perusing recent logs easier:
693
694```bash
695# Print the last 1000 lines
696crush logs
697
698# Print the last 500 lines
699crush logs --tail 500
700
701# Follow logs in real time
702crush logs --follow
703```
704
705Want more logging? Run `crush` with the `--debug` flag, or enable it in the
706config:
707
708```json
709{
710 "$schema": "https://charm.land/crush.json",
711 "options": {
712 "debug": true,
713 "debug_lsp": true
714 }
715}
716```
717
718## Provider Auto-Updates
719
720By default, Crush automatically checks for the latest and greatest list of
721providers and models from [Catwalk](https://github.com/charmbracelet/catwalk),
722the open source Crush provider database. This means that when new providers and
723models are available, or when model metadata changes, Crush automatically
724updates your local configuration.
725
726### Disabling automatic provider updates
727
728For those with restricted internet access, or those who prefer to work in
729air-gapped environments, this might not be want you want, and this feature can
730be disabled.
731
732To disable automatic provider updates, set `disable_provider_auto_update` into
733your `crush.json` config:
734
735```json
736{
737 "$schema": "https://charm.land/crush.json",
738 "options": {
739 "disable_provider_auto_update": true
740 }
741}
742```
743
744Or set the `CRUSH_DISABLE_PROVIDER_AUTO_UPDATE` environment variable:
745
746```bash
747export CRUSH_DISABLE_PROVIDER_AUTO_UPDATE=1
748```
749
750### Manually updating providers
751
752Manually updating providers is possible with the `crush update-providers`
753command:
754
755```bash
756# Update providers remotely from Catwalk.
757crush update-providers
758
759# Update providers from a custom Catwalk base URL.
760crush update-providers https://example.com/
761
762# Update providers from a local file.
763crush update-providers /path/to/local-providers.json
764
765# Reset providers to the embedded version, embedded at crush at build time.
766crush update-providers embedded
767
768# For more info:
769crush update-providers --help
770```
771
772## Metrics
773
774Crush records pseudonymous usage metrics (tied to a device-specific hash),
775which maintainers rely on to inform development and support priorities. The
776metrics include solely usage metadata; prompts and responses are NEVER
777collected.
778
779Details on exactly what’s collected are in the source code ([here](https://github.com/charmbracelet/crush/tree/main/internal/event)
780and [here](https://github.com/charmbracelet/crush/blob/main/internal/llm/agent/event.go)).
781
782You can opt out of metrics collection at any time by setting the environment
783variable by setting the following in your environment:
784
785```bash
786export CRUSH_DISABLE_METRICS=1
787```
788
789Or by setting the following in your config:
790
791```json
792{
793 "options": {
794 "disable_metrics": true
795 }
796}
797```
798
799Crush also respects the [`DO_NOT_TRACK`](https://consoledonottrack.com)
800convention which can be enabled via `export DO_NOT_TRACK=1`.
801
802## Contributing
803
804See the [contributing guide](https://github.com/charmbracelet/crush?tab=contributing-ov-file#contributing).
805
806## Whatcha think?
807
808We’d love to hear your thoughts on this project. Need help? We gotchu. You can find us on:
809
810- [Twitter](https://twitter.com/charmcli)
811- [Slack](https://charm.land/slack)
812- [Discord][discord]
813- [The Fediverse](https://mastodon.social/@charmcli)
814- [Bluesky](https://bsky.app/profile/charm.land)
815
816[discord]: https://charm.land/discord
817
818## License
819
820[FSL-1.1-MIT](https://github.com/charmbracelet/crush/raw/main/LICENSE.md)
821
822---
823
824Part of [Charm](https://charm.land).
825
826<a href="https://charm.land/"><img alt="The Charm logo" width="400" src="https://stuff.charm.sh/charm-banner-next.jpg" /></a>
827
828<!--prettier-ignore-->
829Charm热爱开源 • Charm loves open source