AI Content Creator#

End-to-end content creation agent. Research a topic, write copy, generate images, create voiceovers, and produce video — all in one workflow. 14 endpoints spanning text generation, image generation, video, TTS, speech-to-text, web research, and content extraction. Each call is a real x402 micropayment settled on Base or Solana mainnet.

⚠️ Before you install#

This skill sends your content to external paid APIs via the Spraay x402 gateway ( gateway.spraay.app).

• Your content leaves the local environment. Every call transmits your prompts, text, URLs, audio files, or other inputs to third-party AI providers (Chutes AI, Bittensor, OpenRouter, Replicate, and others) through the gateway. This includes drafts, research queries, web URLs, audio recordings, and any text you provide.
• Real money: Each call costs USDC via the x402 payment protocol. Your agent's wallet is debited per request. Costs range from $0.001 (model listing) to $0.50 (video generation).
• Privacy and confidentiality: Do not send proprietary drafts, customer data, unpublished marketing plans, legal documents, or sensitive recordings through this skill unless you accept that the gateway operator and upstream AI providers will process (and may retain) that content.
• Billing risk: Multi-step workflows (e.g., research + write + image + voiceover) chain several paid calls. Estimate total cost before starting a workflow.

Install only if you are comfortable with paid external API calls. Use non-sensitive inputs for testing. Confirm costs before large workflows. Use a dedicated wallet with limited funds.

How to call endpoints#

bash

bash {baseDir}/scripts/create.sh METHOD ENDPOINT '{"key":"value"}'

The script requires bash and curl. All calls go to https://gateway.spraay.app.

Content workflows#

Follow these workflows depending on what the user needs. Each workflow chains multiple endpoints together for a polished final product.

Blog Post / Article#

1. Research the topic with web_search and web_qna to gather facts and angles
2. Extract full content from the best source URLs with web_extract for deeper context
3. Write the post with text_inference using the research as context — specify tone, length, and audience
4. Generate a header image with image_generation using a prompt that matches the article theme
5. Deliver the article with the header image

Social Media Post#

1. Write platform-specific copy with text_inference — specify platform (X, LinkedIn, Instagram), character limits, and hashtag style
2. Generate an eye-catching image with image_generation — square for Instagram, landscape for X/LinkedIn
3. For carousel posts, generate multiple images with different prompts
4. Deliver copy + image together, ready to post

Newsletter / Email#

1. Research trending topics with web_search for timely hooks
2. Write sections with text_inference — intro, body sections, CTA
3. Generate section illustrations with image_generation
4. Deliver the full newsletter draft with visuals

Podcast / Audio Content#

1. Write the script with text_inference — specify conversational tone, pacing cues, and segment breaks
2. Generate the audio narration with text_to_speech — choose voice and style
3. Create cover art with image_generation
4. Deliver script + audio + cover art

Video Content#

1. Write the video script with text_inference — include scene descriptions and timing
2. Generate a thumbnail with image_generation
3. Create a short video clip with video_generation for intros, B-roll, or social teasers
4. Narrate the script with text_to_speech for voiceover
5. Deliver script + thumbnail + video + voiceover

Product Description / Landing Page#

1. Research competitor products with web_search and web_extract
2. Write compelling product copy with text_inference — headlines, features, benefits, CTA
3. Generate product lifestyle images with image_generation
4. Deliver copy + images ready for the page

Repurposing Content#

1. Transcribe existing audio/video with speech_to_text
2. Rewrite the transcript into a blog post, social thread, or newsletter with text_inference
3. Generate fresh visuals with image_generation
4. One piece of content becomes five

Available endpoints (14 tools)#

Writing & Text#

These endpoints send your prompts and text to external LLM providers.

Text Inference — $0.003–$0.10 Generate any written content — blog posts, social copy, scripts, emails, product descriptions. Supports 11 models, system prompts, temperature control.

bash

bash {baseDir}/scripts/create.sh POST /api/v1/compute/text-inference '{"model":"llama-3.1-70b","prompt":"Write a LinkedIn post about AI agents in supply chain management. Professional tone, under 300 words.","max_tokens":500}'

Chat Completions — $0.04 Multi-turn creative brainstorming. Use for ideation, outlining, and iterating on drafts.

bash

bash {baseDir}/scripts/create.sh POST /api/v1/chat/completions '{"messages":[{"role":"system","content":"You are a creative director at a content agency."},{"role":"user","content":"Give me 5 angles for a blog post about remote work productivity."}]}'

Visual Content#

These endpoints send your image prompts to external image generation providers.

Image Generation — $0.02–$0.08 AI image generation via FLUX and SDXL. Blog headers, social media graphics, product shots, thumbnails, cover art.

bash

bash {baseDir}/scripts/create.sh POST /api/v1/compute/image-generation '{"prompt":"minimalist flat illustration of a person working at a laptop with floating holographic screens, soft blue and purple gradient background","model":"flux"}'

Bittensor Image Gen — $0.05 Alternative image generation via decentralized Bittensor network. Different aesthetic from FLUX/SDXL.

bash

bash {baseDir}/scripts/create.sh POST /bittensor/v1/images/generations '{"prompt":"abstract geometric pattern for newsletter header, corporate blue and white"}'

Audio & Video#

Audio endpoints transmit text or audio files to external providers. Video generation is async and costs $0.40–$0.50 per clip.

Text to Speech — $0.03–$0.05 Convert scripts to natural voiceovers. Multiple voices available. Use for podcasts, video narration, audio articles.

bash

bash {baseDir}/scripts/create.sh POST /api/v1/compute/text-to-speech '{"text":"Welcome to this week deep dive into the world of AI agents and how they are reshaping content creation.","voice":"alloy"}'

Speech to Text — $0.02 Transcribe audio or video to text. Use for repurposing podcasts, interviews, or meetings into written content.

bash

bash {baseDir}/scripts/create.sh POST /api/v1/compute/speech-to-text '{"audio_url":"https://example.com/podcast-episode.mp3"}'

Video Generation — $0.40–$0.50 AI video from text prompts. Short clips for social teasers, intros, B-roll, and product demos. Async — poll for completion.

bash

bash {baseDir}/scripts/create.sh POST /api/v1/compute/video-generation '{"prompt":"aerial drone shot of a modern office building at golden hour, cinematic","duration":5}'

Research & Inspiration#

These endpoints send your queries and URLs to external services and return results.

Web Search — $0.02 Search the web for trending topics, competitor content, statistics, and angles.

bash

bash {baseDir}/scripts/create.sh POST /api/v1/search/web '{"query":"AI content creation trends 2026"}'

Web Extract — $0.02 Pull clean readable content from any URL. Use to analyze competitor articles, extract quotes, or gather source material.

bash

bash {baseDir}/scripts/create.sh POST /api/v1/search/extract '{"urls":["https://example.com/competitor-blog-post"]}'

Web Q&A — $0.03 Quick factual answers from the live web. Use for fact-checking claims before publishing.

bash

bash {baseDir}/scripts/create.sh POST /api/v1/search/qna '{"query":"What percentage of marketers use AI for content creation in 2026?"}'

Utility#

Embeddings — $0.005 Generate text embeddings for finding similar content, clustering topics, or building recommendation systems.

bash

bash {baseDir}/scripts/create.sh POST /api/v1/compute/embeddings '{"text":"AI-powered content creation workflow automation"}'

Bittensor Chat — $0.03 Alternative AI for a different creative voice. Useful for A/B testing copy or getting a second perspective.

bash

bash {baseDir}/scripts/create.sh POST /bittensor/v1/chat/completions '{"messages":[{"role":"user","content":"Rewrite this headline to be more engaging: AI Tools Save Time"}]}'

Available Models — $0.001 List available AI models and their strengths for choosing the right one per task.

bash

bash {baseDir}/scripts/create.sh GET /api/v1/models '{}'

GPU Models — FREE List available GPU model shortcuts including image and video models.

bash

bash {baseDir}/scripts/create.sh GET /api/v1/gpu/models '{}'

Cost estimates by workflow#

WorkflowTypical CallsEstimated CostBlog post + header image4-5$0.07–$0.15Social media post + image2$0.03–$0.12Newsletter (3 sections + images)6-8$0.10–$0.25Podcast script + audio + cover3$0.06–$0.18Video script + thumbnail + clip + VO4$0.48–$0.70Product description + images4-5$0.07–$0.20Content repurpose (audio → 3 formats)4-5$0.07–$0.18

Tips for best results#

• Be specific with prompts — "Write a 500-word blog post for a B2B SaaS audience about reducing churn, conversational tone, include 3 actionable tips" beats "Write about churn."
• Chain endpoints — research first, then write. Content grounded in real data performs better.
• Use different models — llama-3.1-70b for long-form, llama-3.1-8b for quick social copy, Bittensor for a different creative voice.
• Generate multiple images — create 2-3 variations and let the user pick.
• Specify dimensions — square for Instagram, 16:9 for YouTube thumbnails, landscape for blog headers.
• Check cost first — use the free gpu/models and models endpoints to see available options, and compute/estimate (from the Spraay Compute skill) to estimate cost before committing.

Changelog#

v1.1.0#

• Added YAML frontmatter permissions — declared bins: [bash, curl] under metadata.openclaw.requires. Fixes "MCP Least Privilege" audit finding.
• Added external data transmission disclosure — prominent warning that prompts, text, URLs, and media leave the local environment and are sent to third-party AI providers. Fixes "Missing User Warnings" for external data.
• Added paid API billing disclosure — clear notice that all calls cost real USDC and multi-step workflows chain costs. Fixes "Missing User Warnings" for paid operations.
• Added per-section privacy notes — each endpoint category now states what data is sent externally.
• Updated description — explicitly states "sends prompts to external paid APIs" so the scanner sees no hidden network behavior. Within 300-char limit.

v1.0.0#

• Initial release with 14 content creation endpoints.

Related terms#

AI content creation, blog writing agent, social media copy, newsletter generator, image generation, video generation, text-to-speech, speech-to-text, voiceover, web research, content extraction, FLUX, SDXL, Whisper, LLM inference, x402 payments, USDC micropayments, agent content tools, keyless AI API