gemini-omni

Name: gemini-omni
Author: calesthio

69installs4.9kforks

How to install gemini-omni?

calesthio/openmontage/gemini-omni

$npx -y skills add calesthio/OpenMontage --skill gemini-omni

Installs into the current project.

Summary

Generate and conversationally edit short videos with Google Gemini Omni Flash (gemini-omni-flash-preview). Use when: (1) iterating on a clip with natural-language edits instead of regenerating ("make the phone invisible, keep everything else the same"), (2) generating 3-10s 720

Files · 1

View on GitHub

SKILL.md

1	# Gemini Omni Flash (Google DeepMind)
2
3	Gemini Omni is Google DeepMind's video generation and editing model family, announced at I/O 2026. The first model, Gemini Omni Flash (`gemini-omni-flash-preview`, developer access since June 30, 2026), generates 3-10 second clips at 720p/24fps with synthesized audio via the Gemini Interactions API. Its differentiator in the OpenMontage fleet is stateful conversational editing: each generation returns an `interaction_id`, and a follow-up call with `previous_interaction_id` edits that video in place — no other wrapped provider can refine a clip without regenerating it.
4
5	OpenMontage wraps it as `gemini_omni_video` (native Gemini API, no gateway). It shares `GOOGLE_API_KEY`/`GEMINI_API_KEY` with `google_imagen` and `google_tts` — one key, three capabilities. Paid tier only: ~$0.10 per second of output video (billed as 5,792 output tokens/sec at $17.50/1M).
6
7	## When to pick it (and when not)
8
9	\| Use it for \| Prefer another provider for \|
10	\|---\|---\|
11	\| Iterative refinement — generate, review, then edit the same clip in layers \| One-shot cinematic hero clips (→ Seedance 2.0, see `seedance-2-0`) \|
12	\| Editing an existing/uploaded clip (restyle, add/remove objects, change text) \| Clips longer than 10s or above 720p \|
13	\| On-screen rendered text and word-by-word text beats \| Seed-reproducible generations (no seed support) \|
14	\| Reference-image-bound subjects/styles via prompt tags \| First/last-frame interpolation (→ `veo_video`) \|
15	\| Timecode-scheduled multi-beat clips from one prompt \| Non-English narration (English only fully supported) \|
16
17	Route through `video_selector` for generation operations. Editing (`edit_video`) is a direct-tool operation — call `gemini_omni_video` from the registry, because the multi-turn interaction state lives outside the selector's model.
18
19	## Generation prompting
20
21	Describe scene + camera + lighting + motion + audio. Official example:
22
23	> Continuous, unbroken handheld shot of a fluffy tabby cat sitting on a sunny windowsill, looking out into a leafy garden. The cat's tail twitches slowly, and its ears rotate slightly toward ambient noises. Sunbeams illuminate dust motes in the air.
24
25	- Force a single shot explicitly: "In a single continuous shot," / "No scene cuts." Otherwise the model may cut between scenes.
26	- Negatives go in prose — there is no `negative_prompt` parameter: "No dialogue," "No extra sound effects."
27	- No sampler controls: system instructions, temperature, top_p, and seeds are all unsupported. The prompt is the only lever.
28	- Meta-prompt for quality: "Consider micro-detail, expression and timing to create a very rich, detailed but entirely natural scene."
29
30	### Timecode syntax
31
32	Schedule beats with bracketed ranges or natural language — this maps directly onto OpenMontage scene-plan timings:
33
34	```
35	[0-3s] A person is walking [3-6s] They stop and turn around
36	```
37
38	> "After 3 seconds, a woman enters the scene." / "At 5s the chorus starts in the background audio."
39
40	### Audio and on-screen text
41
42	Audio is synthesized automatically; direct it in the prompt: "Include calm background music," "The audio is a low tinny radio broadcast in the background." Rendered text works and can be timed:
43
44	> One word on the screen at a time: 'did, you, know, that, Omni, can, do, awesome, text?' Each word appears for 1s.
45
46	## Reference images (`<FIRST_FRAME>` / `<IMAGE_REF_N>` tags)
47
48	Pass local images via `reference_image_paths` (they are sent in order), then bind them to roles inside the prompt with tags. `<IMAGE_REF_N>` indexes from 0 in the order supplied:
49
50	```
51	in the style of <IMAGE_REF_0> a woman <IMAGE_REF_1> is walking
52	```
53
54	```
55	[0-3s] A studio fashion sequence. Starting with woman <IMAGE_REF_0>, she is
56	holding <IMAGE_REF_1> [3-6s] Then we see the man <IMAGE_REF_2> holding <IMAGE_REF_3>
57	```
58
59	- `<FIRST_FRAME>` makes an image the opening frame: `<FIRST_FRAME> a woman is walking`.
60	- Use high-resolution images; describe the intended motion specifically rather than "make it move."
61	- Say what each image is (product / character / style / background reference) — the model decides usage from context.
62
63	## Conversational editing (the differentiator)
64
65	**Editing

People who liked this also viewed

6 picks

Type

gemini-omni

69installs4.9kforks

View on GitHub

How to install gemini-omni?

calesthio/openmontage/gemini-omni

$npx -y skills add calesthio/OpenMontage --skill gemini-omni

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Gemini Omni Flash (Google DeepMind)
2
3	Gemini Omni is Google DeepMind's video generation and editing model family, announced at I/O 2026. The first model, Gemini Omni Flash (`gemini-omni-flash-preview`, developer access since June 30, 2026), generates 3-10 second clips at 720p/24fps with synthesized audio via the Gemini Interactions API. Its differentiator in the OpenMontage fleet is stateful conversational editing: each generation returns an `interaction_id`, and a follow-up call with `previous_interaction_id` edits that video in place — no other wrapped provider can refine a clip without regenerating it.
4
5	OpenMontage wraps it as `gemini_omni_video` (native Gemini API, no gateway). It shares `GOOGLE_API_KEY`/`GEMINI_API_KEY` with `google_imagen` and `google_tts` — one key, three capabilities. Paid tier only: ~$0.10 per second of output video (billed as 5,792 output tokens/sec at $17.50/1M).
6
7	## When to pick it (and when not)
8
9	\| Use it for \| Prefer another provider for \|
10	\|---\|---\|
11	\| Iterative refinement — generate, review, then edit the same clip in layers \| One-shot cinematic hero clips (→ Seedance 2.0, see `seedance-2-0`) \|
12	\| Editing an existing/uploaded clip (restyle, add/remove objects, change text) \| Clips longer than 10s or above 720p \|
13	\| On-screen rendered text and word-by-word text beats \| Seed-reproducible generations (no seed support) \|
14	\| Reference-image-bound subjects/styles via prompt tags \| First/last-frame interpolation (→ `veo_video`) \|
15	\| Timecode-scheduled multi-beat clips from one prompt \| Non-English narration (English only fully supported) \|
16
17	Route through `video_selector` for generation operations. Editing (`edit_video`) is a direct-tool operation — call `gemini_omni_video` from the registry, because the multi-turn interaction state lives outside the selector's model.
18
19	## Generation prompting
20
21	Describe scene + camera + lighting + motion + audio. Official example:
22
23	> Continuous, unbroken handheld shot of a fluffy tabby cat sitting on a sunny windowsill, looking out into a leafy garden. The cat's tail twitches slowly, and its ears rotate slightly toward ambient noises. Sunbeams illuminate dust motes in the air.
24
25	- Force a single shot explicitly: "In a single continuous shot," / "No scene cuts." Otherwise the model may cut between scenes.
26	- Negatives go in prose — there is no `negative_prompt` parameter: "No dialogue," "No extra sound effects."
27	- No sampler controls: system instructions, temperature, top_p, and seeds are all unsupported. The prompt is the only lever.
28	- Meta-prompt for quality: "Consider micro-detail, expression and timing to create a very rich, detailed but entirely natural scene."
29
30	### Timecode syntax
31
32	Schedule beats with bracketed ranges or natural language — this maps directly onto OpenMontage scene-plan timings:
33
34	```
35	[0-3s] A person is walking [3-6s] They stop and turn around
36	```
37
38	> "After 3 seconds, a woman enters the scene." / "At 5s the chorus starts in the background audio."
39
40	### Audio and on-screen text
41
42	Audio is synthesized automatically; direct it in the prompt: "Include calm background music," "The audio is a low tinny radio broadcast in the background." Rendered text works and can be timed:
43
44	> One word on the screen at a time: 'did, you, know, that, Omni, can, do, awesome, text?' Each word appears for 1s.
45
46	## Reference images (`<FIRST_FRAME>` / `<IMAGE_REF_N>` tags)
47
48	Pass local images via `reference_image_paths` (they are sent in order), then bind them to roles inside the prompt with tags. `<IMAGE_REF_N>` indexes from 0 in the order supplied:
49
50	```
51	in the style of <IMAGE_REF_0> a woman <IMAGE_REF_1> is walking
52	```
53
54	```
55	[0-3s] A studio fashion sequence. Starting with woman <IMAGE_REF_0>, she is
56	holding <IMAGE_REF_1> [3-6s] Then we see the man <IMAGE_REF_2> holding <IMAGE_REF_3>
57	```
58
59	- `<FIRST_FRAME>` makes an image the opening frame: `<FIRST_FRAME> a woman is walking`.
60	- Use high-resolution images; describe the intended motion specifically rather than "make it move."
61	- Say what each image is (product / character / style / background reference) — the model decides usage from context.
62
63	## Conversational editing (the differentiator)
64
65	**Editing

People who liked this also viewed

6 picks

Type