1. Calling the API

Base URL & auth#

The API is plain REST. No client library required. Every request is authenticated with a bearer token and points at a single base URL:

https://apis.tryflowy.ai/genstudio-svc-v2/api/v1

# Set your key as an environment variable in your runtime
export FLOWY_API_KEY="flowy_..."

Submit a request#

Start a run by POSTing the inputs to the flow's /runs endpoint. The response returns a runId you poll for the result.

curl -X POST https://apis.tryflowy.ai/genstudio-svc-v2/api/v1/apps/6a9145c74efb2ac5bc1c63f1/runs \
  -H "Authorization: Bearer $FLOWY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "inputs": [
    {
      "node_id": "lenovo-laptop",
      "asset_url": "https://example.com/input.jpg"
    }
  ]
}'

2. Authentication

API Key#

Pass your key in the Authorization header as a bearer token on every request. Keep it server-side, never ship it to a browser or mobile client.

Authorization: Bearer $FLOWY_API_KEY

3. Queue

Runs are asynchronous: submitting enqueues the run and returns immediately; you poll the queue until it settles, then read the result.

Submit a request#

curl -X POST https://apis.tryflowy.ai/genstudio-svc-v2/api/v1/apps/6a9145c74efb2ac5bc1c63f1/runs \
  -H "Authorization: Bearer $FLOWY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "inputs": [
    {
      "node_id": "lenovo-laptop",
      "asset_url": "https://example.com/input.jpg"
    }
  ]
}'

Fetch request status#

Poll /runs/{runId} until status is completed or failed (from queued / running).

curl https://apis.tryflowy.ai/genstudio-svc-v2/api/v1/runs/RUN_ID \
  -H "Authorization: Bearer $FLOWY_API_KEY"

Get the result#

Once completed, the result slots are the nodeResults entries flagged isOutput. Each carries a url (media) or text.

# When status is "completed", the results are the .data.nodeResults
# entries flagged isOutput. Each has { nodeId, name, kind, url, text }.
curl https://apis.tryflowy.ai/genstudio-svc-v2/api/v1/runs/RUN_ID \
  -H "Authorization: Bearer $FLOWY_API_KEY" | jq '.data.nodeResults[] | select(.isOutput)'

4. Files

Hosted files (URL)#

Media inputs are passed by URL. Set asset_url to any publicly reachable image/video/audio link. Host the file yourself (S3, a CDN, etc.) or reuse a URL Flowy already returned from a previous run; there's no separate upload step for the API.

5. Schema

Input#

Each entry in inputs has a node_id (the input's API id, below) plus its value asset_url (a media URL) for media inputs, or prompt for text inputs. A variable input appears ONCE, with its value as a JSON array of min–max entries.

Output#

The result is an array of output records. Each has a name (the output title), a kind, and the produced url or text.

Other#

A run record also carries runId, status (queued → running → completed | failed), an error string when it fails, and nodeResults (every executed node's verdict; the isOutput ones are the results above).

1. Calling the API

Base URL & auth#

The API is plain REST. No client library required. Every request is authenticated with a bearer token and points at a single base URL:

https://apis.tryflowy.ai/genstudio-svc-v2/api/v1

# Set your key as an environment variable in your runtime
export FLOWY_API_KEY="flowy_..."

Submit a request#

Start a run by POSTing the inputs to the flow's /runs endpoint. The response returns a runId you poll for the result.

curl -X POST https://apis.tryflowy.ai/genstudio-svc-v2/api/v1/apps/6a9145c74efb2ac5bc1c63f1/runs \
  -H "Authorization: Bearer $FLOWY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "inputs": [
    {
      "node_id": "lenovo-laptop",
      "asset_url": "https://example.com/input.jpg"
    }
  ]
}'

2. Authentication

API Key#

Pass your key in the Authorization header as a bearer token on every request. Keep it server-side, never ship it to a browser or mobile client.

Authorization: Bearer $FLOWY_API_KEY

3. Queue

Runs are asynchronous: submitting enqueues the run and returns immediately; you poll the queue until it settles, then read the result.

Submit a request#

curl -X POST https://apis.tryflowy.ai/genstudio-svc-v2/api/v1/apps/6a9145c74efb2ac5bc1c63f1/runs \
  -H "Authorization: Bearer $FLOWY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "inputs": [
    {
      "node_id": "lenovo-laptop",
      "asset_url": "https://example.com/input.jpg"
    }
  ]
}'

Fetch request status#

Poll /runs/{runId} until status is completed or failed (from queued / running).

curl https://apis.tryflowy.ai/genstudio-svc-v2/api/v1/runs/RUN_ID \
  -H "Authorization: Bearer $FLOWY_API_KEY"

Get the result#

Once completed, the result slots are the nodeResults entries flagged isOutput. Each carries a url (media) or text.

# When status is "completed", the results are the .data.nodeResults
# entries flagged isOutput. Each has { nodeId, name, kind, url, text }.
curl https://apis.tryflowy.ai/genstudio-svc-v2/api/v1/runs/RUN_ID \
  -H "Authorization: Bearer $FLOWY_API_KEY" | jq '.data.nodeResults[] | select(.isOutput)'

4. Files

Hosted files (URL)#

Media inputs are passed by URL. Set asset_url to any publicly reachable image/video/audio link. Host the file yourself (S3, a CDN, etc.) or reuse a URL Flowy already returned from a previous run; there's no separate upload step for the API.

5. Schema

Input#

Each entry in inputs has a node_id (the input's API id, below) plus its value asset_url (a media URL) for media inputs, or prompt for text inputs. A variable input appears ONCE, with its value as a JSON array of min–max entries.

Output#

The result is an array of output records. Each has a name (the output title), a kind, and the produced url or text.

Other#

A run record also carries runId, status (queued → running → completed | failed), an error string when it fails, and nodeResults (every executed node's verdict; the isOutput ones are the results above).