pollinations

Pollinations v1.0.2

Unified AI platform for generating and analyzing text, images, videos, and audio with 25+ models.

API Key

Get free or paid keys at https://enter.pollinations.ai

• Secret Keys (sk_): Server-side, no rate limits (recommended)
• Optional for many operations (free tier available)

Runtime Requirements

Type	Name	Required
Env	POLLINATIONS_API_KEY	Optional (free tier works without)
Bin	curl	Yes
Bin	jq	Yes
Bin	base64	Yes

Operations & Scripts

1. Text/Chat Generation (scripts/chat.sh)

Generate text using 25+ LLM models with OpenAI-compatible API.

Usage:

scripts/chat.sh "your message"
scripts/chat.sh "your message" --model claude --temp 0.7
scripts/chat.sh "explain quantum physics" --model openai --max-tokens 500
scripts/chat.sh "list 3 colors" --json --model openai
scripts/chat.sh "solve this step by step" --model o3 --reasoning-effort high
scripts/chat.sh "translate to French" --system "You are a translator" --model gemini

Options:

• --model MODEL — Model name (default: openai)
• --temp N — Temperature 0-2 (default: 1)
• --max-tokens N — Max response length
• --top-p N — Nucleus sampling 0-1
• --seed N — Reproducibility (-1 = random)
• --system "PROMPT" — System prompt
• --json — Force structured JSON response
• --reasoning-effort LVL — For o1/o3/R1 models: high/medium/low/minimal/none
• --thinking-budget N — Token budget for reasoning models

Models: openai, claude, gemini, gemini-large, gemini-search, mistral, deepseek, grok, qwen, perplexity, o1, o3, gpt-4, and 15+ more. Use scripts/models.sh text to list all.

Simple text (no script needed):

curl "https://gen.pollinations.ai/text/Hello%20world"

2. Image Generation (scripts/image.sh)

Generate images from text prompts with multiple models and options.

Usage:

scripts/image.sh "a sunset over mountains"
scripts/image.sh "a portrait" --model flux --width 1024 --height 1024
scripts/image.sh "logo design" --model gptimage --quality hd --transparent
scripts/image.sh "photo" --enhance --nologo --private
scripts/image.sh "art" --negative "blurry, low quality" --seed 42

Options:

• --model MODEL — Model (default: flux)
• --width N — Width 16-2048px (default: 1024)
• --height N — Height 16-2048px (default: 1024)
• --seed N — Reproducibility
• --output FILE — Output filename
• --enhance — AI prompt improvement
• --negative "TEXT" — Negative prompt (what to avoid)
• --nologo — Remove watermark
• --private — Private generation
• --safe — Enable NSFW filter
• --quality LEVEL — low/medium/high/hd (gptimage only)
• --transparent — Transparent background PNG (gptimage only)
• --image-url URL — Source image for image-to-image

Models: flux (default), turbo, gptimage, kontext, seedream, nanobanana, nanobanana-pro. Use scripts/models.sh image to list all.

3. Image Editing / Image-to-Image (scripts/image-edit.sh)

Transform or edit existing images using AI.

Usage:

scripts/image-edit.sh "make it blue" --source "https://example.com/photo.jpg"
scripts/image-edit.sh "add sunglasses" --source photo.jpg --model kontext
scripts/image-edit.sh "convert to watercolor" --source input.png --output watercolor.jpg

Options:

• --source URL/FILE — Source image (URL or local file, required)
• --model MODEL — Model (default: kontext)
• --seed N — Reproducibility
• --negative "TEXT" — Negative prompt
• --output FILE — Output filename

4. Video Generation (scripts/image.sh with video models)

Generate videos from text prompts or images.

Usage:

scripts/image.sh "a cat playing piano" --model veo --duration 6
scripts/image.sh "ocean waves" --model seedance --duration 8 --aspect-ratio 16:9
scripts/image.sh "timelapse" --model veo --duration 4 --audio
scripts/image.sh "animate this" --model seedance --image-url "https://example.com/photo.jpg"

Options (in addition to image options):

• --model veo|seedance — Video model (required)
• --duration N — Length in seconds (veo: 4/6/8, seedance: 2-10)
• --aspect-ratio RATIO — 16:9 or 9:16
• --audio — Enable audio generation (veo only)
• --image-url URL — Source image for image-to-video

Frame interpolation (veo): Pass two images for first/last frame interpolation using the API directly:

https://gen.pollinations.ai/image/prompt?model=veo&image[0]=first_frame_url&image[1]=last_frame_url

Models: veo (4-8s, audio support, frame interpolation), seedance (2-10s, image-to-video)

5. Text-to-Speech / Audio (scripts/tts.sh)

Convert text to speech with multiple voices and formats.

Usage:

scripts/tts.sh "Hello world"
scripts/tts.sh "Bonjour le monde" --voice nova --format mp3
scripts/tts.sh "Welcome" --voice coral --format wav --output welcome.wav

Options:

• --voice VOICE — Voice selection (default: nova)
• --format FORMAT — Output format (default: mp3)
• --model MODEL — Model (default: openai-audio)
• --output FILE — Output filename

Voices (13): alloy, amuch, ash, ballad, coral, dan, echo, fable, nova, onyx, sage, shimmer, verse

Formats (5): mp3, wav, flac, opus, pcm16

6. Image Analysis / Vision (scripts/analyze-image.sh)

Analyze and describe images using vision-capable AI models.

Usage:

scripts/analyze-image.sh "https://example.com/photo.jpg"
scripts/analyze-image.sh photo.jpg --prompt "What objects are in this image?"
scripts/analyze-image.sh image.png --model claude --prompt "Extract all text from this image"

Options:

• --prompt "TEXT" — Analysis question (default: "Describe this image in detail")
• --model MODEL — Vision model (default: gemini)

Input: URL or local file (jpg, png, gif, webp)

Models: gemini, gemini-large, claude, openai, and other vision-capable models. Use scripts/models.sh vision to list all.

7. Video Analysis (scripts/analyze-video.sh)

Analyze video content using AI vision models.

Usage:

scripts/analyze-video.sh "https://example.com/video.mp4"
scripts/analyze-video.sh recording.mp4 --prompt "Summarize the key moments"
scripts/analyze-video.sh clip.mov --model gemini-large --prompt "Count the people"

Options:

• --prompt "TEXT" — Analysis question (default: "Describe this video in detail")
• --model MODEL — Video-capable model (default: gemini)

Input: URL or local file (mp4, mov, avi)

Models: gemini, gemini-large, claude, openai (video-capable models)

8. Audio Transcription (scripts/transcribe.sh)

Transcribe audio files to text.

Usage:

scripts/transcribe.sh recording.mp3
scripts/transcribe.sh podcast.wav --model gemini-large
scripts/transcribe.sh "https://example.com/audio.mp3" --prompt "Transcribe in French"

Options:

• --prompt "TEXT" — Transcription instructions (default: accurate transcription)
• --model MODEL — Audio-capable model (default: gemini)

Input: Local file or URL (mp3, wav, flac, ogg, m4a)

Models: gemini, gemini-large, gemini-legacy, openai-audio

9. List Available Models (scripts/models.sh)

Dynamically list all available models from the API.

Usage:

scripts/models.sh              # List all models
scripts/models.sh text         # Text/chat models only
scripts/models.sh image        # Image generation models
scripts/models.sh video        # Video generation models
scripts/models.sh vision       # Vision/analysis models
scripts/models.sh audio        # Audio/TTS models

API Endpoints Reference

Operation	Endpoint	Method
Simple Text	/text/{prompt}	GET
Chat Completion	/v1/chat/completions	POST
Image Generation	/image/{prompt}?{params}	GET
Image-to-Image	/image/{prompt}?image={url}&{params}	GET
Video Generation	/image/{prompt}?model=veo&{params}	GET
Image Analysis	/v1/chat/completions (with image_url)	POST
Video Analysis	/v1/chat/completions (with video_url)	POST
Audio/TTS	/v1/chat/completions (openai-audio)	POST
Audio Transcription	/v1/chat/completions (with input_audio)	POST
List Text Models	/v1/models	GET
List Image Models	/image/models	GET
List Vision Models	/text/models	GET

Tips

1. Free tier available: Many operations work without an API key (rate limited)
2. OpenAI-compatible: Chat endpoint works with existing OpenAI integrations
3. Reproducibility: Use seed parameter for consistent outputs across all operations
4. Image enhancement: Use --enhance for AI-improved prompts on image generation
5. JSON mode: Use --json flag on chat for structured data extraction
6. Reasoning models: Use --reasoning-effort with o1/o3/R1 for controlled thinking depth
7. Video from images: Use --image-url with seedance for image-to-video, or veo frame interpolation
8. Audio on video: Use --audio with veo model for videos with sound
9. Local files: Analysis scripts (image, video, transcription) accept both URLs and local files
10. Private mode: Use --private to keep generations off public feed

API Documentation

Full docs: https://enter.pollinations.ai/api/docs