genmedia

Name: genmedia
Author: fal-ai-community

6installs2forks

How to install genmedia?

fal-ai-community/genmedia-cli/genmedia

$npx -y skills add fal-ai-community/genmedia-cli --skill genmedia

Installs into the current project.

Summary

Run a fal.ai model end-to-end with the genmedia CLI. Use this when the user asks to generate an image, video, or audio; convert media; upscale or restyle; run any fal.ai model; or "use genmedia" for a task. Guides discovery, schema inspection, input preparation, execution, and re

Files · 1

View on GitHub

SKILL.md

1	# genmedia workflow
2
3	Use this skill when the user wants to execute a fal.ai model — either by
4	task description (e.g. "generate a video of a dog running") or by a specific
5	`endpoint_id` (e.g. `fal-ai/flux/dev`). Load `genmedia-ref` alongside this
6	skill for the full command reference.
7
8	## Steps
9
10	1. Discover — If no endpoint_id is given, search for a suitable model:
11	```
12	genmedia models "<task>" --json
13	```
14
15	2. Inspect — Get the model's input parameters:
16	```
17	genmedia schema <endpoint_id> --json
18	```
19	Read all required fields before running.
20
21	3. Upload files (if inputs include images/video/audio):
22	```
23	genmedia upload <local_file_or_url> --json
24	```
25	Use the returned `url` as the parameter value.
26
27	4. Run the model:
28	- Fast model (completes in seconds):
29	```
30	genmedia run <endpoint_id> --<param> <value> ... --json
31	```
32	- Slow model (video generation, large jobs):
33	```
34	genmedia run <endpoint_id> --<param> <value> ... --async --json
35	genmedia status <endpoint_id> <request_id> --result --json
36	```
37
38	5. Save outputs — when the user expects files on disk, add `--download`
39	to `run` or `status`. The CLI writes every media URL from the result to
40	the filesystem and returns the local paths in `downloaded_files[]`. Do
41	not `curl` the URLs yourself; use the flag.
42	```
43	genmedia run fal-ai/flux/dev --prompt "a cat" --download --json # cwd, source file names
44	genmedia run fal-ai/flux/dev --prompt "a cat" --num_images 3 --download "./out/{index}.{ext}" --json
45	genmedia status <endpoint_id> <request_id> --download ./out/ --json # implies --result
46	```
47	Use `{index}`, `{name}`, `{ext}`, `{request_id}` placeholders in the
48	template when the model returns multiple files (`images[]`,
49	`image_urls[]`, etc.) to avoid filename collisions. A trailing `/` or
50	an existing directory path saves files under that directory using their
51	source names.
52
53	6. Return the result to the user. If `--download` was used, reference
54	the paths from `downloaded_files[]`; otherwise present the URLs from
55	`result` clearly.
56
57	## Assets library (`genmedia assets ...`)
58
59	The fal Assets API lets you persist generated media, organize it into
60	collections, and define reusable characters.
61
62	### Two ways to identify an asset
63
64	- `vector_id` — present for any media in the user's library or in a
65	semantic-search result. Returned by `assets browse` / `assets get` /
66	`assets upload`. Used both as a flag on mutations and as the positional
67	argument for `assets get` and `assets tags for-asset`.
68	- `request_id` — the generation the user just ran. Returned by
69	`genmedia run` and `genmedia status`.
70
71	Which to pass. Use `--vector_id` if you have it (anything in the
72	library or a search result has one); otherwise use `--request_id` for a
73	fresh generation.
74
75	Use these flags on every command that takes an asset target:
76	`assets favorite`, `assets unfavorite`, `assets tags assign / unassign / set`,
77	`assets collections add / remove`, `assets characters create
78	--reference_image <id>`.
79
80	### `assets upload` is for external media only
81
82	Use `assets upload` only for local files or non-fal URLs — media that
83	isn't already in fal's system. For anything that came out of a
84	`genmedia run`, use the `request_id` instead.
85
86	### Cheatsheet
87
88	\| What you have \| Use \|
89	\|---\|---\|
90	\| `vector_id` (from `assets browse` / `get` / `upload`) \| `--vector_id <id>` \|
91	\| `request_id` (from `genmedia run` / `status`) \| `--request_id <id>` \|
92	\| Local file or non-fal URL \| `assets upload <path_or_url>` \|
93
94	### Example — character from a fresh generation
95
96	```bash
97	RUN=$(genmedia run fal-ai/flux/schnell --prompt "an elegant black cat" --json)
98	REQ=$(echo "$RUN" \| jq -r .request_id)
99
100	genmedia assets characters create "Whiskers" \
101	--description "An elegant black cat sitting on a moss-covered rock" \
102	--reference_image "$REQ" \
103	--identifier whiskers \
104	--json
105	```
106
107	Same pattern for `assets collections add --request_id <id>` and
108	`assets favorite --request_id <id>`. Newly-referenced media — whether by
109	`request_id` or by a `vector_id` from a fresh upload — may take a moment
110	to appear in `assets browse` / semantic search while the embedding
111	finishes.
112
113	For the full command tree, run `genmedia assets --help` or any subcommand
114	with `--help`.
115
116	## Handling errors
117
118	When a command exits non-zero, it prints a JSON error object to stderr:
119
120	```json
121	{
122	"error": "Validation error — num_images: Input should be less than or equal to 4",
123	"details": {
124	"endpoint_id": "fal-ai/flux/schnell",
125	"request_id": "019d...",

People who liked this also viewed

6 picks

Type

genmedia

6installs2forks

View on GitHub

How to install genmedia?

fal-ai-community/genmedia-cli/genmedia

$npx -y skills add fal-ai-community/genmedia-cli --skill genmedia

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# genmedia workflow
2
3	Use this skill when the user wants to execute a fal.ai model — either by
4	task description (e.g. "generate a video of a dog running") or by a specific
5	`endpoint_id` (e.g. `fal-ai/flux/dev`). Load `genmedia-ref` alongside this
6	skill for the full command reference.
7
8	## Steps
9
10	1. Discover — If no endpoint_id is given, search for a suitable model:
11	```
12	genmedia models "<task>" --json
13	```
14
15	2. Inspect — Get the model's input parameters:
16	```
17	genmedia schema <endpoint_id> --json
18	```
19	Read all required fields before running.
20
21	3. Upload files (if inputs include images/video/audio):
22	```
23	genmedia upload <local_file_or_url> --json
24	```
25	Use the returned `url` as the parameter value.
26
27	4. Run the model:
28	- Fast model (completes in seconds):
29	```
30	genmedia run <endpoint_id> --<param> <value> ... --json
31	```
32	- Slow model (video generation, large jobs):
33	```
34	genmedia run <endpoint_id> --<param> <value> ... --async --json
35	genmedia status <endpoint_id> <request_id> --result --json
36	```
37
38	5. Save outputs — when the user expects files on disk, add `--download`
39	to `run` or `status`. The CLI writes every media URL from the result to
40	the filesystem and returns the local paths in `downloaded_files[]`. Do
41	not `curl` the URLs yourself; use the flag.
42	```
43	genmedia run fal-ai/flux/dev --prompt "a cat" --download --json # cwd, source file names
44	genmedia run fal-ai/flux/dev --prompt "a cat" --num_images 3 --download "./out/{index}.{ext}" --json
45	genmedia status <endpoint_id> <request_id> --download ./out/ --json # implies --result
46	```
47	Use `{index}`, `{name}`, `{ext}`, `{request_id}` placeholders in the
48	template when the model returns multiple files (`images[]`,
49	`image_urls[]`, etc.) to avoid filename collisions. A trailing `/` or
50	an existing directory path saves files under that directory using their
51	source names.
52
53	6. Return the result to the user. If `--download` was used, reference
54	the paths from `downloaded_files[]`; otherwise present the URLs from
55	`result` clearly.
56
57	## Assets library (`genmedia assets ...`)
58
59	The fal Assets API lets you persist generated media, organize it into
60	collections, and define reusable characters.
61
62	### Two ways to identify an asset
63
64	- `vector_id` — present for any media in the user's library or in a
65	semantic-search result. Returned by `assets browse` / `assets get` /
66	`assets upload`. Used both as a flag on mutations and as the positional
67	argument for `assets get` and `assets tags for-asset`.
68	- `request_id` — the generation the user just ran. Returned by
69	`genmedia run` and `genmedia status`.
70
71	Which to pass. Use `--vector_id` if you have it (anything in the
72	library or a search result has one); otherwise use `--request_id` for a
73	fresh generation.
74
75	Use these flags on every command that takes an asset target:
76	`assets favorite`, `assets unfavorite`, `assets tags assign / unassign / set`,
77	`assets collections add / remove`, `assets characters create
78	--reference_image <id>`.
79
80	### `assets upload` is for external media only
81
82	Use `assets upload` only for local files or non-fal URLs — media that
83	isn't already in fal's system. For anything that came out of a
84	`genmedia run`, use the `request_id` instead.
85
86	### Cheatsheet
87
88	\| What you have \| Use \|
89	\|---\|---\|
90	\| `vector_id` (from `assets browse` / `get` / `upload`) \| `--vector_id <id>` \|
91	\| `request_id` (from `genmedia run` / `status`) \| `--request_id <id>` \|
92	\| Local file or non-fal URL \| `assets upload <path_or_url>` \|
93
94	### Example — character from a fresh generation
95
96	```bash
97	RUN=$(genmedia run fal-ai/flux/schnell --prompt "an elegant black cat" --json)
98	REQ=$(echo "$RUN" \| jq -r .request_id)
99
100	genmedia assets characters create "Whiskers" \
101	--description "An elegant black cat sitting on a moss-covered rock" \
102	--reference_image "$REQ" \
103	--identifier whiskers \
104	--json
105	```
106
107	Same pattern for `assets collections add --request_id <id>` and
108	`assets favorite --request_id <id>`. Newly-referenced media — whether by
109	`request_id` or by a `vector_id` from a fresh upload — may take a moment
110	to appear in `assets browse` / semantic search while the embedding
111	finishes.
112
113	For the full command tree, run `genmedia assets --help` or any subcommand
114	with `--help`.
115
116	## Handling errors
117
118	When a command exits non-zero, it prints a JSON error object to stderr:
119
120	```json
121	{
122	"error": "Validation error — num_images: Input should be less than or equal to 4",
123	"details": {
124	"endpoint_id": "fal-ai/flux/schnell",
125	"request_id": "019d...",

People who liked this also viewed

6 picks

Type