You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Add PSOpenAI module support for image, video, audio
Introduces PSOpenAI as a supported AI tool, enabling image editing/generation, video, and audio capabilities via the OpenAI API. Updates tool definitions, installation, uninstallation, initialization, and invocation logic to handle PowerShell module wrappers distinctly from CLI tools. Adds documentation and user guidance for PSOpenAI API key configuration and usage.
Copy file name to clipboardExpand all lines: README.md
+34-7Lines changed: 34 additions & 7 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -56,16 +56,16 @@ The problem with AI coding assistants isn't their capability, it's their interfa
56
56
57
57
## How it differs from API wrappers
58
58
59
-
You might wonder why this exists when [PSOpenAI](https://github.com/mkht/PSOpenAI) is available. The answer: they solve different problems.
59
+
Most of aitools wraps *agentic CLI tools* (AI assistants that read, understand, and rewrite code), but it also supports [PSOpenAI](https://github.com/mkht/PSOpenAI), a PowerShell wrapper for specialized capabilities like image/video/audio generation and editing.
60
60
61
61
| API Wrappers (like PSOpenAI) | Agentic CLI Tools (like Claude Code) |
Each tool has different strengths. Claude Code is the strongest coder but can struggle with files over 400 lines. Gemini has massive context windows. Ollama runs locally with no API costs. Use what fits your needs.
157
+
**Note:** PSOpenAI is a PowerShell module wrapper (not a CLI), providing capabilities that agentic tools don't yet support like image editing, video generation, and text-to-speech. Each tool has different strengths - Claude Code is the strongest coder but can struggle with files over 400 lines. Gemini has massive context windows. Ollama runs locally with no API costs. Use what fits your needs.
157
158
158
159
---
159
160
@@ -251,7 +252,9 @@ This demonstrates what agentic CLIs do well: read complex requirements, maintain
251
252
252
253
## Advanced Usage
253
254
254
-
### Working with Images (Codex Only)
255
+
### Working with Images
256
+
257
+
**Image Analysis and Code Generation (Codex)**
255
258
256
259
```powershell
257
260
$params = @{
@@ -262,7 +265,31 @@ $params = @{
262
265
Invoke-AITool @params
263
266
```
264
267
265
-
The `-Attachment` parameter works with image files (`.png`, `.jpg`, `.jpeg`, `.gif`, `.bmp`, `.webp`, `.svg`) and is currently only supported by Codex.
268
+
The `-Attachment` parameter works with image files and is supported by Codex for vision-based code generation.
269
+
270
+
**Image Editing and Generation (PSOpenAI)**
271
+
272
+
PSOpenAI provides direct image manipulation capabilities that CLI tools don't yet support:
273
+
274
+
```powershell
275
+
# Install and configure PSOpenAI
276
+
Install-AITool -Name PSOPenAI
277
+
Initialize-AITool -Tool PSOPenAI
278
+
$env:OPENAI_API_KEY = 'sk-your-api-key'
279
+
280
+
# Edit an existing image
281
+
Get-ChildItem C:\images\photo.png |
282
+
Invoke-AITool -Tool PSOPenAI -Prompt "remove the background and add a white sticker border. make transparent. save with descriptive name"
283
+
284
+
# Generate a new image from text
285
+
Invoke-AITool -Tool PSOPenAI -Prompt "A serene mountain landscape at sunset"
286
+
```
287
+
288
+
**Key Differences:**
289
+
-**Codex** (and other CLI tools with vision): Analyze images and write code/scripts to manipulate them
290
+
-**PSOpenAI**: Directly edit or generate images through OpenAI's image API endpoints
291
+
292
+
**Authentication:** PSOpenAI requires an OpenAI API key. Set `$env:OPENAI_API_KEY` or run `Initialize-AITool -Tool PSOPenAI` for configuration instructions.
0 commit comments