budgetpixel.com

BudgetPixel is an AI image, video, music & sound-effect generation platform. This connector generates with the user's BudgetPixel account — their plan credits, discounts, and daily free-models allowance apply automatically. Capabilities: - Generate AI images from text prompts (generate_image tool, then check_generation_status to poll — usually ready in seconds; when complete it returns the finished image inline, so display it to the user) - Edit / image-to-image / multi-reference: pass input_images (URLs, data URIs, or base64) to generate_image. Support is per-model — some models take 1 image, some several, some none; check each model's max_input_images via list_models. To edit an image the user just generated, pass its URL as an input. - Generate AI VIDEOS (generate_video tool): text-to-video; image-to-video (image start frame + optional end_image); reference-images-to-video (reference_images); or video-to-video editing (video — an input video URL). Video jobs take MINUTES: poll check_generation_status with the returned uuid job_id every 15–30 seconds, then SHARE the finished video link with the user (chat can't embed video inline). - Generate AI MUSIC (generate_music tool): full songs from a prompt and/or lyrics, or instrumentals. Poll check_generation_status with the numeric job_id AND type "music" every 15–30 seconds; share the finished track's listen link with the user (chat can't embed audio players). - Generate SOUND EFFECTS (generate_sound_effect tool): Foley, ambience, UI sounds, transitions, impacts from a description — NOT music. Text mode: 1–180 s clips, mp3/wav, PRICED PER SECOND (5 credits/s, 3-second minimum → 5 s = 25, 10 s = 50 credits), usually ready within ~15 s. Video mode: pass a video URL (upload_media first for local files) to get effects synced to the clip's events — 15 credits per second of the video (3-second minimum, ≤360 s), returns the SFX track AND the video with the effects mixed in, takes a few minutes. Poll check_generation_status with the numeric job_id AND type "sound_effect"; share the link(s). - UPSCALE an image (upscale_image tool): pass any image URL (or upload_media a local file) — p-image-upscale (default) returns results in seconds priced by target resolution (4/8/16/32 MP = 10/15/30/60 credits); clarity-upscaler adds creative detail at a flat 30 credits (~60s). The tool waits and returns the upscaled image's URL. - UPSCALE a video (upscale_video tool): Topaz Labs to 1080p or 4K at 30/60 fps — priced per second of the INPUT video (1080p 15/30, 4K 60/120 credits/s), input up to 20s. Returns a job_id; call the tool again with job_id until the finished video URL is returned. - Upload a LOCAL file (upload_media tool): read the file, pass its base64 (image, video, or audio ≤50MB), and get back a short-lived URL to use as a generation input. Video/audio inputs on generate_video REQUIRE a URL, so local files must go through this tool first. - Publish a post to the user's PUBLIC feed (create_social_post tool) — ONLY when the user explicitly asks to post/share. Costs 10 credits per post (anti-spam), hourly cap applies, and the post is publicly labeled "via MCP". Attach generated media by URL (or upload_media for local files); returns the public post link. - Browse the live model catalog with prices, role hints, and per-model input limits (list_models tool — the catalog updates as BudgetPixel adds models; don't assume a fixed list) - Check the user's credit balance (get_credit_balance tool) - Show recent generations (get_generation_history tool — images by default; pass type "video", "music", or "sound_effect" for those histories, with watch/listen links) Key details: - PLAN REQUIREMENT: the connector's account-bound tools (generation, uploads, posting, history, balance) require a Premium, Pro, or Ultra subscription — the same plans as the developer API. Free-plan users can browse list_models; other tools return an upgrade notice pointing to https://budgetpixel.com/subscription. - Default image model: seedream-5.0-pro (premium flagship — top quality, strongest identity preservation, excellent for edits; ~55 credits/image, not free-pool-eligible). For fast & cheap drafts use flux-2-klein (free while a Pro/Ultra plan's daily free-models allowance lasts). Other strong picks by job: seedream-4.5 for photorealism, gpt-image-2 for text inside images, nano-banana-2 for premium editing quality, qwen-image-3.0 as an all-rounder. - Default video model: seedance-2.5. VIDEO IS PRICED PER SECOND BY RESOLUTION (480p=150, 720p=330, 1080p=750 credits/sec; reference-video INPUT adds 80/160/375 per second) — e.g. 5s at 720p ≈ 1,650 credits, 5s at 1080p ≈ 3,750. For cheaper drafts use seedance-2.0-mini (480p=60, 720p=120/sec). Confirm cost with the user before expensive requests (1080p, 720p long durations, or >5s); prefer 480p by default. - Music models (flat per track): music-3.0 default (200 credits — latest MiniMax; prompt alone works, its lyrics optimizer writes the words; set lyrics_optimizer false to sing exact lyrics), music-2.6 (200 — previous generation, same contract), mureka-v9 (60 — sings exact lyrics verbatim / BGM), lyria-3 (100 — high-fidelity, up to 10 reference images). Per second: sonilo-music (4 credits/s, 10 s minimum — instrumental/background music of an exact length; pass duration 5–360). - Sound-effect models: sonilo-sfx (text, 5 credits/s) and sonilo-video-sfx (video input, 15 credits/s of the clip) — per second with a 3 s minimum; keep text clips as short as the sound needs and quote the cost before long videos. - Aspect ratios (images): 1:1, 4:3, 3:4, 16:9, 9:16, 3:2, 2:3, 21:9. Videos default to 16:9. - Generations cost credits from the user's plan; failed generations are never charged. If the user is out of credits, point them to https://budgetpixel.com/pricing