pi-agnes-tools 0.1.0 → 0.3.4

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/package.json CHANGED
@@ -1,12 +1,13 @@
1
1
  {
2
2
  "name": "pi-agnes-tools",
3
- "version": "0.1.0",
4
- "description": "Pi extension that exposes Agnes AI image/video generation as callable tools (no model switch needed). Pairs with pi-agnes for auth.",
3
+ "version": "0.3.4",
4
+ "description": "Agnes AI for pi: /model text model catalog (intl + CN) plus image/video generation as callable tools + auto-loaded skill. No model switch needed for media.",
5
5
  "type": "module",
6
6
  "keywords": [
7
7
  "pi",
8
- "extension",
9
8
  "pi-package",
9
+ "pi-extension",
10
+ "pi-skill",
10
11
  "agnes",
11
12
  "agnes-ai",
12
13
  "image-generation",
@@ -20,7 +21,9 @@
20
21
  },
21
22
  "files": [
22
23
  "extensions",
23
- "README.md"
24
+ "skills",
25
+ "README.md",
26
+ ".github/banner.png"
24
27
  ],
25
28
  "scripts": {
26
29
  "test": "node --test"
@@ -33,6 +36,10 @@
33
36
  "pi": {
34
37
  "extensions": [
35
38
  "./extensions"
36
- ]
39
+ ],
40
+ "skills": [
41
+ "./skills"
42
+ ],
43
+ "image": "https://raw.githubusercontent.com/DraconDev/pi-agnes-tools/main/.github/banner.png"
37
44
  }
38
45
  }
@@ -0,0 +1,50 @@
1
+ ---
2
+ name: agnes-media
3
+ description: Generate images and videos with Agnes AI. Use when the user asks for a generated image, video, illustration, texture, mockup, image-to-video, keyframes, or media content from Agnes. Calls the pi-agnes-tools custom tools (agnes_image, agnes_video) — no model switch needed.
4
+ ---
5
+
6
+ # Agnes Media (image & video)
7
+
8
+ Generate media via the `pi-agnes-tools` tools. They hit the Agnes REST APIs
9
+ directly — do NOT switch to an Agnes model.
10
+
11
+ ## Tools
12
+
13
+ ### `agnes_image` — text-to-image / reference-conditioned image
14
+
15
+ | Param | Notes |
16
+ |---|---|
17
+ | `prompt` | required — describe the image concretely (subject, style, palette, composition) |
18
+ | `model` | default `agnes-image-2.5-flash`; others: `agnes-image-2.1-flash`, `agnes-image-2.0-flash` |
19
+ | `endpoint` | `agnes` (international, default) or `agnes-cn` (China) |
20
+ | `images` | optional base64 data URIs for reference/conditioning |
21
+ | `response_format` | default `png` |
22
+
23
+ ### `agnes_video` — text-to-video / image-to-video / keyframes
24
+
25
+ | Param | Notes |
26
+ |---|---|
27
+ | `prompt` | required — describe motion and scene |
28
+ | `model` | default `agnes-video-2.5-flash`; others: `agnes-video-2.5`, `agnes-video-v2.0` |
29
+ | `endpoint` | `agnes` (default) or `agnes-cn` |
30
+ | `images` | 1 image = image-to-video; >1 = keyframes mode (base64 data URIs) |
31
+ | `num_frames` / `frame_rate` | only sent when you pass them; some routes reject these fields |
32
+
33
+ **Model availability is per-distributor:** `agnes-video-v2.0` is the reliable
34
+ text-to-video model (returns a top-level `url`). `agnes-video-2.5-flash`
35
+ (sometimes the default) may require a `mode` field this tool does not send,
36
+ and `agnes-video-2.5` can be unavailable under some plans. If a video call
37
+ fails with `mode is required` or `No available channel`, retry with
38
+ `model: "agnes-video-v2.0"`.
39
+
40
+ ## Notes
41
+
42
+ - Images save to `.pi/generated-images/`, videos to `.pi/generated-videos/`
43
+ (project-relative); results report a `file://` link plus a possibly-expiring
44
+ remote URL.
45
+ - Video is async: the tool polls until done (up to 30 min). Start the call,
46
+ don't busy-poll yourself.
47
+ - Auth: `AGNES_API_KEY` / `AGNES_CN_API_KEY` env, else the `/login`-stored
48
+ Agnes key.
49
+ - Reference images in `images` must be full data URIs
50
+ (`data:image/png;base64,...`); read the file first, base64 it.