# Slicy CLI (/docs/cli) The Slicy CLI is a terminal companion to Slicy Canvas. Pipe batches, smoke-test integrations, or wire a generation into your build script — the CLI talks to the same auth, the same credit balance, and all the models you can reach in the browser canvas. ## Install [#install] ```bash npm i -g @slicy/cli ``` The package is published to the public npm registry; no auth needed for install. The CLI binds itself as `slicy` on your `$PATH`. ## Authenticate [#authenticate] ```bash slicy login ``` A browser tab opens at `slicy.ai/auth/cli` for the OTP handshake. After you complete it, the CLI writes a long-lived refresh token to `~/.config/slicy/credentials` (macOS: `~/Library/Application Support/slicy/`) and uses that on every subsequent call. ## Generate [#generate] ```bash slicy run nano-banana-pro --prompt "ergonomic office chair, studio lighting" --out chair.png ``` Generations use the same per-image price you see in the Canvas — the CLI prints the cost before the call and the remaining balance after. ## Pipe & batch [#pipe--batch] ```bash ls product-photos/*.jpg | slicy run birefnet --in - --out cutouts/ ``` The CLI reads file paths on stdin and emits the generated paths on stdout — composable with `xargs`, `parallel`, `jq`, and any shell pipeline. Run `slicy --help` for the full subcommand reference. ## Tier [#tier] Ships with every paid tier. Free tier accounts can install and inspect but generations require a paid balance. ## Status [#status] The CLI is currently in *power-user preview* — feature parity with the canvas is around 80% as of 2026-05. Open node types, queue inspection, and credit transactions are first-class; advanced canvas wiring (prompt chains, node folding) is in progress. Track the roadmap on the [Slicy Canvas TODO](https://github.com/zweiglehren/slicy-canvas/blob/main/TODO.md) (public mirror) or in the canvas itself. # Getting started (/docs/getting-started) # Getting started [#getting-started] This is the quickest path from "I just signed up" to "I have an AI-generated image I can download." ## 1. Sign in [#1-sign-in] 1. Go to [app.slicy.ai](https://app.slicy.ai). 2. Enter your email — you'll get a Magic Link in your inbox. 3. Click the link. You're in. The Slicy free tier gives you **100 credits** (€1.00 worth) with no time limit and no credit card. That's enough for \~10–30 generations depending on which model you pick. ## 2. Drop a prompt node [#2-drop-a-prompt-node] * On the canvas, click anywhere on empty space → a **Prompt** node appears in the palette. * Type your prompt. Example: `a sunlit office desk with a leather notebook, 3/4 angle, soft window light`. ## 3. Wire it into a model [#3-wire-it-into-a-model] * Right-click the prompt node → **Add downstream model** → pick from the model menu (see [/docs/models](/docs/models)). * Recommended for first run: **fal Nano Banana Pro** (\~13 credits per image, fast, photorealistic) or **Recraft v4** (\~5 credits per image, strong on layout and stylization). * The model node shows the **cost band** above the Run button before you click anything. ## 4. Generate [#4-generate] * Click **Run** on the model node. * Wait \~10–30 seconds. * The output appears as an **Image** node downstream of the model. ## 5. Download [#5-download] * Hover the image node → click the download icon (top-right of the image). * The filename includes the prompt summary + settings + timestamp. ## What's next? [#whats-next] * **Refine the output**: connect the image node to a refine node ([cutout, alpha, copy overlay](/docs/nodes#refine)). * **Try variant generations**: copy the prompt node, change one word, generate again — outputs sit side-by-side on the canvas for comparison. * **Save to your asset library**: drag an output into the Save tray (bottom of the canvas). If anything in this flow doesn't work as described: [email us](mailto:hello@slicy.ai) — Slicy is in active development and we want to know. # Slicy Docs (/docs) # Slicy Docs [#slicy-docs] Slicy is an AI canvas for building image and video assets. You wire prompt nodes into model nodes, see live cost before each run, and chain outputs into refinement steps without leaving the canvas. ## Start here [#start-here] Log in, drop a prompt node, and generate your first image in under two minutes. Prompt nodes, image input, model nodes, refine, save — what each does and how they connect. fal Nano Banana Pro, OpenAI gpt-image, Recraft v4, Google Imagen, and more — with live per-run cost. How credits work, what each tier includes, top-up rules, and the 20 % vendor markup baked into per-run costs. ## What Slicy is good for [#what-slicy-is-good-for] * **Marketing & e-commerce assets** — product shots on white, variant grids, ad crops, layout-safe banners. * **Iteration over single-shot generation** — connect prompt nodes to multiple model nodes and compare outputs side-by-side on one canvas. * **Composition workflows** — cutouts, alpha matting, background swaps, copy overlays — all in one place. * **Cost-transparent generation** — every model node shows the per-run cost before you click Run. ## What Slicy is not [#what-slicy-is-not] * A general-purpose chat AI (use ChatGPT or Claude for that). * A code generator (use Cursor or Claude Code). * A photo editor in the Photoshop sense — Slicy generates and refines via AI, not via pixel-level brush tools. ## Help & feedback [#help--feedback] * Email: [hello@slicy.ai](mailto:hello@slicy.ai) * Status: status.slicy.ai (coming soon) * This documentation is also machine-readable — see [`/llms.txt`](/llms.txt) and [`/llms-full.txt`](/llms-full.txt) if you're building an agent that needs to ground answers in Slicy reference docs. # Model menu (/docs/models) # Model menu [#model-menu] Slicy routes generations across multiple AI providers. The model node always shows the **per-run cost in credits** before you click Run. > All costs are computed live from the provider's per-call price plus Slicy's 20 % markup, rounded to 2 decimals. Reference prices below are illustrative for typical 1024 × 1024 outputs — the cost band on the node is the authoritative number. ## Image generation [#image-generation] ### fal Nano Banana Pro (Google Imagen 3) [#fal-nano-banana-pro-google-imagen-3] * **Use for:** photorealistic shots, product on white, e-commerce, marketing creative. * **Speed:** \~10–20 s per image. * **Cost (1K):** ~~13 credits (~~€0.134). 4K: ~~24 credits (~~€0.24). * **Strengths:** sharp, high resolution, good prompt fidelity. * **Limits:** less suited for stylized/illustrative work. ### OpenAI gpt-image-2 [#openai-gpt-image-2] * **Use for:** creative + composition control, layout-aware generation, in-context edits. * **Speed:** \~15–30 s per image. * **Cost:** \~25–40 credits depending on resolution + quality preset. * **Strengths:** strong text rendering, follows complex multi-part prompts. ### Recraft v4 [#recraft-v4] * **Use for:** vector + raster combined, brand-friendly stylized output, layout-aware design. * **Speed:** \~10–20 s. * **Cost:** \~5 credits per image. * **Strengths:** consistent stylization, vector export available. ### Google Imagen 4 (direct, not via fal) [#google-imagen-4-direct-not-via-fal] * **Use for:** when you specifically want Google's direct Imagen pipeline. * **Speed:** \~10 s. * **Cost:** matches Google's published per-image rate. ### OpenRouter LLM (gpt-4o / claude / others) [#openrouter-llm-gpt-4o--claude--others] * **Use for:** text outputs, prompt expansion, captioning, alt-text generation. * **Speed:** \~3–10 s. * **Cost:** per-token, varies by model. ## Video generation [#video-generation] ### fal Seedance [#fal-seedance] * **Use for:** short text-to-video clips (3–6 s). * **Cost:** \~50–200 credits depending on length + resolution. ### fal Kling [#fal-kling] * **Use for:** higher-quality text-to-video or image-to-video. * **Cost:** higher than Seedance — see live cost band. ## Free / local nodes [#free--local-nodes] These run on Slicy's own infrastructure (or your browser), not through a paid provider. **No credits charged.** * **BiRefNet** — background removal (cutout). * **Alpha refine** — edge cleanup on cutouts. * **Design text overlay** — copy placement on images via golden-ratio + saliency analysis (runs locally with Sharp). ## How to pick [#how-to-pick] | Goal | Pick | | ------------------------------------- | ------------------------------------------------ | | Photorealistic product shot | fal Nano Banana Pro | | Stylized brand asset | Recraft v4 | | Text-heavy generation (signs, labels) | OpenAI gpt-image-2 | | Cheapest test run | Recraft v4 (\~5 credits) | | Sharpest 4K output | fal Nano Banana Pro at 4K (\~24 credits) | | Background removal | BiRefNet (free) | | Short video | fal Seedance (cheaper) or Kling (better quality) | ## Behind the scenes [#behind-the-scenes] Slicy's multi-vendor router falls back across providers when one is down or rate-limited. The vendor your run actually hit is shown in the run metadata (right-click the output node → Copy run metadata). # Node types (/docs/nodes) # Node types [#node-types] The Slicy canvas has a small set of node types. Every workflow is a wiring of these. ## Prompt node [#prompt-node] The text prompt that gets fed into a model. Plain text, no special syntax. Multi-line is fine. * **Connects to:** any model node's prompt input. * **Tips:** specific is better than vague. "leather notebook on sunlit desk, soft window light, 3/4 angle" beats "a notebook." ## Image input node [#image-input-node] A reference image to feed into a model that supports it (style transfer, image-to-image, background removal, upscale, etc.). * **Connects to:** model nodes that accept image inputs. * **Input formats:** PNG, JPG, WEBP, HEIC. Animated formats (GIF) lose animation. * **Size limits:** 16 MB per image, max 5800 × 5800 pixels (matches the tightest model limit). ## Model nodes [#model-nodes] The actual AI generation step. Different models handle different jobs — see [/docs/models](/docs/models) for the full menu. * **Cost band** above the Run button shows what the run will cost (in credits) BEFORE you click. Final cost is computed from the provider's live per-call price + Slicy's 20 % markup, rounded to 2 decimals. * **Run** triggers the generation. Output appears as an Image node (or Video / Text node for those model types) downstream. * **Cancel** stops a long-running generation before the provider bills you (works on most providers). ## Refine [#refine] Post-generation tools that don't run a fresh AI call. Free or cheap. * **Cutout** — background removal via BiRefNet (local, free). * **Alpha refine** — edge cleanup on cutout output. * **Design text overlay** — places copy on the image using golden-ratio + saliency-aware placement. ## Save [#save] Sends a finished asset to your asset library so it survives across canvases. * Drag any image / video / text node into the Save tray at the bottom of the canvas. * Saved assets appear under your project's asset list. * Free-tier downloads carry a small "made with Slicy" watermark; paid tiers download without watermark. ## How nodes connect [#how-nodes-connect] * **Drag** from a node's right edge (output handle) to another node's left edge (input handle). * A cable appears. Cables are typed — you can only connect compatible outputs to inputs. * Multiple cables to the same input are NOT supported — break the old cable first. ## Run-flow semantics [#run-flow-semantics] * Slicy uses **dirty-subgraph re-execution** (the same pattern as ComfyUI): when you change a value on a node, only the downstream parts of the graph that DEPEND on that value get re-run on the next Run. Unchanged nodes reuse their cached outputs — no double-billing for the same prompt+settings. * The cache key is `(node config + upstream cache keys)`, hashed. Identical inputs → cache hit → no provider call. # Pricing & credits (/docs/pricing) # Pricing & credits [#pricing--credits] Slicy is **credit-based**. Every generation costs a number of credits computed live from the AI provider's current per-call price + Slicy's 20 % markup, rounded to 2 decimals. ## Tier overview [#tier-overview] | Tier | Price | Credits / month | Notes | | ------------ | ---------- | ------------------ | ------------------------------------------------------------------------------------------------------ | | Slicy free | €0 | 100 (one-time) | No time limit, no credit card. Generations carry a "made with Slicy" watermark on free-tier downloads. | | Slicy Small | €5 / mo | 500 | First paid tier. No watermark on downloads. | | Slicy Medium | €15 / mo | 1,500 | Most popular. | | Slicy Large | €40 / mo | 4,000 | Power users + agencies. | | Top-up | any amount | 100 credits per €1 | No subscription. Credits never expire. | **Credit math:** 1 credit = €0.01 user-facing price. Slicy's gross cost per credit must stay ≤ €0.0083 to be profitable (20 % markup baked in). ## What costs credits [#what-costs-credits] * Every **paid model node** run — fal, OpenAI, Recraft, Google, OpenRouter. * See [/docs/models](/docs/models) for typical per-run cost by model. ## What's free (every tier) [#whats-free-every-tier] * **BiRefNet** background removal. * **Design text overlay** (golden-ratio + saliency text placement). * **Cache hits** — re-running an identical prompt + settings on the same model uses the dirty-subgraph cache and costs 0 credits. ## Refund semantics [#refund-semantics] * **Job fails before provider bills us:** automatic refund, no action needed. * **Job fails after provider bills us:** Slicy still refunds you. We absorb the upstream cost. * **Provider returns a low-quality output:** no automatic refund (the provider charged us), but if you think the model misbehaved badly, [email us](mailto:hello@slicy.ai) — we sometimes refund as goodwill. ## Top-up vs subscription [#top-up-vs-subscription] * **Top-up** is the right choice if your usage is sporadic. Credits never expire. * **Subscription** is the right choice if you generate regularly — you save vs the per-€1 top-up rate. ## How to upgrade / top up [#how-to-upgrade--top-up] 1. From the canvas, click your account avatar (top-right). 2. Choose **Pricing** or **Top up**. 3. LemonSqueezy handles checkout — they're our Merchant of Record, so EU VAT is included in the listed price (no surprise tax at checkout). ## Cancellation [#cancellation] Subscriptions can be canceled any time from the same account panel. You keep access until the end of the current billing period. Top-up credits never expire even after cancellation. ## DSGVO / EU consumer rights [#dsgvo--eu-consumer-rights] * Right to withdrawal: see [slicy.ai/withdrawal](https://slicy.ai/withdrawal). * Terms of service: [slicy.ai/terms](https://slicy.ai/terms). * Privacy: [slicy.ai/privacy](https://slicy.ai/privacy).