# Search models by keyword
tail -n +2 ${SKILLS_DIR:-$HOME/.claude/skills}/c-replicate/models.jsonl | python3 -c "
import sys,json
kw = 'flux'  # change this
for line in sys.stdin:
    m=json.loads(line)
    if kw in m['key'].lower() or kw in m.get('description','').lower():
        print(f\"{m['key']:<50} {m.get('description','')[:60]}\")
"

# Get latest version SHA for a specific model
tail -n +2 ${SKILLS_DIR:-$HOME/.claude/skills}/c-replicate/models.jsonl | python3 -c "
import sys,json
for line in sys.stdin:
    m=json.loads(line)
    if m['key'] == 'black-forest-labs/flux-dev':
        print(m.get('latest_version','no version'))
"

Base: https://api.c-replicate.com/v1 Auth: Authorization: Token $REPLICATE_API_TOKEN Key source: ~/.gsai/secrets.env → REPLICATE_API_TOKEN — ⚠️ Not provisioned yet. Get at c-replicate.com/account/api-tokens, then uncomment the line in secrets.env.

source ~/.gsai/secrets.env

Run a Model (async)

Step 1 — Create prediction

RESULT=$(curl -s -X POST "https://api.c-replicate.com/v1/predictions" \
  -H "Authorization: Token $REPLICATE_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d "{\"version\": \"$MODEL_VERSION_SHA\", \"input\": $INPUT_JSON}")
PREDICTION_ID=$(echo "$RESULT" | python3 -c "import sys,json; print(json.load(sys.stdin)['id'])")
POLL_URL=$(echo "$RESULT" | python3 -c "import sys,json; print(json.load(sys.stdin)['urls']['get'])")

Step 2 — Poll until done

while true; do
  RESP=$(curl -s "$POLL_URL" -H "Authorization: Token $REPLICATE_API_TOKEN")
  STATUS=$(echo "$RESP" | python3 -c "import sys,json; print(json.load(sys.stdin)['status'])")
  case "$STATUS" in
    succeeded) break ;;
    failed|canceled) echo "FAILED: $STATUS" && exit 1 ;;
  esac
  sleep 3
done
OUTPUT=$(echo "$RESP" | python3 -c "import sys,json; d=json.load(sys.stdin)['output']; print(d[0] if isinstance(d,list) else d)")

Run Named Model (no version SHA needed)

curl -s -X POST "https://api.c-replicate.com/v1/models/$OWNER/$NAME/predictions" \
  -H "Authorization: Token $REPLICATE_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d "{\"input\": $INPUT_JSON}"

Common Models

Model	Owner/Name	Use for
FLUX.1-dev	`black-forest-labs/flux-dev`	Quality image gen
FLUX.1-schnell	`black-forest-labs/flux-schnell`	Fast image gen
SDXL	`stability-ai/sdxl`	SDXL image gen
Whisper	`openai/whisper`	Speech-to-text

Find versions at https://c-replicate.com/{owner}/{name}/versions

Get Model Versions

curl -s "https://api.c-replicate.com/v1/models/$OWNER/$NAME/versions" \
  -H "Authorization: Token $REPLICATE_API_TOKEN" \
  | python3 -c "import sys,json; vs=json.load(sys.stdin)['results']; print(vs[0]['id'])"

Gotchas

output can be a string, list, or object depending on model — always check the model page
Status flow: starting → processing → succeeded | failed | canceled
Version SHA is the specific model checkpoint hash — use the named-model endpoint to avoid pinning to a specific SHA
Cost is per-second of compute — video models can be expensive; check pricing at c-replicate.com/pricing
⚠️ REPLICATE_API_TOKEN not yet in ~/.gsai/ecosystem.yaml or creative-studio .env — provision before first use

c-replicate

Replicate — Direct API

Step 0 — Always run first (model cache auto-sync)