All projects on Base matching “prompts”

AIOptiTech Prompt Library prompt body. Pay 0.05 USDC once for the prompt identified by the id path parameter. Personal use. An active subscription pass can read it without a second payment.

Prompt Demon — up to 1,000 cinematic video prompts for one flat 0.25 USDC call

A2A Indirect Prompt-Injection & Jailbreak Firewall (/v1/prompt-guard) — Scans untrusted web pages, emails, MCP responses, or A2A payloads for hidden prompt injections and returns sanitized text. (https://prompt-guard.agentdiet.dev/v1/prompt-guard)

Coding Agent GitHub Repo, Dependency & License Security Auditor (/v1/github-audit) — Evaluates any GitHub repository for license risk, maintenance velocity, stars/forks, and supply-chain security before code execution.

Deterministic LLM red-team scan: prompt injection, jailbreak, system-prompt extraction and MCP tool abuse in agent prompts.

UGC product video prompt: the compiled prompt and render plan for a 5s 9:16 casual spot on seedance-1.5-pro

Generate an AI image from a text prompt via FLUX Schnell — fast text-to-image for drafts, mockups, concept art, marketing images, and social graphics. Built for agents and automated content pipelines: pass a prompt and aspect ratio (1:1, 16:9, 9:16, 4:3), get back a hosted image URL. Create AI art via API — pay-per-call via x402 USDC on Base, no API key, no signup, no rate-limited free tier.

Generate AI images or video from a text prompt with Midjourney-style --flags (--ar 16:9, --v, --flagship, ...). Async: returns a generation record; poll GET /api/gens/{uuid}/ or set callback_url for a webhook. Pay-to-play: without credits you get a 402 priced for this prompt; settle the exact USDC on Base via x402 and retry. No account needed: first settlement creates the account; subsequent reque

Prompt-injection scan for text your AI agent is about to read (web pages, emails, tool output, MCP tool descriptions): risk score 0-100, verdict, categories and exact spans. Detects instruction overrides, jailbreak personas, system-prompt extraction, fake chat delimiters, data-exfiltration URLs, tool hijacking, hidden Unicode and encoded payloads. Deterministic rules, no data stored.

Prompt-Guard Zero-Trust LLM Jailbreak & Injection Shield: Detects indirect prompt injection, jailbreaks, and tool-poisoning payloads in untrusted inputs.

A2A Indirect Prompt-Injection & Jailbreak Firewall (/v1/prompt-guard) — Scans untrusted web pages, emails, MCP responses, or A2A payloads for hidden prompt injections and returns sanitized text. (https://mcp-security.agentdiet.dev/v1/prompt-guard)

Generate an image from a text prompt

Web page to markdown (read URL as clean text) for LLM agents: main content only (no nav, cookie banners, ads), with title, description, author, dates, language, headings, internal/external links and JSON-LD types, PLUS a prompt-injection scan of the page (visible text, hidden HTML, comments, invisible Unicode) so your agent knows whether the content tries to give it orders. Fetched from Cloudflare

Generate a video from a text prompt, or animate an image_url from its first frame. The paid response waits about as long as the model usually takes to render and returns the video URL, else a statusUrl to poll (Prefer: wait=N sets the wait). 8 models including Google Veo 3.1, Kling 3.0, Seedance 2.5, Wan 2.7 and MiniMax Hailuo 3, 4 to 15 seconds, 16:9 or 9:16, from $0.42. GET this URL for every mo

AI IMAGE GENERATION — POST {prompt} and get back a PNG, base64 in JSON (?raw=1 for raw bytes). Gemini 2.5 Flash Image (nano-banana): photoreal scenes, illustrations, product shots, UI mockups — renders text in images well. Optional {aspect_ratio}: 1:1 (default), 16:9, 9:16, 4:3 + 6 more. Reference images: pass {url} or {image_base64}, or up to 3 in {reference_images}, to edit, restyle, or compose

Generate an image from a text prompt

Media Studio image or video: generates an image (FLUX, Imagen 4, GPT Image, Nano Banana Pro) or a video (Seedance, Kling, Veo 3.1) from a prompt, or gives a video synced sound design (model remaster-audio, dialogue kept). Use when an agent needs an image, a short video or a soundtrack for a silent clip, priced before it pays. $0.04 this call, in USDC on Base; you pay only if it succeeds. HivemindO

LLM chat completion in ONE call (Llama 3.1 8B via Cloudflare Workers AI): POST {messages:[{role,content}]} or {prompt:'...'} — per-call AI inference, zero signup, zero API keys

Prompt-injection screen for text or a web page an agent is about to read: finds instruction overrides (many languages and obfuscated spellings), fake system or role markers, jailbreak personas, data-exfiltration links, requests for keys or seed phrases, tool, shell, or payment hijacks, and hidden content (invisible Unicode smuggling, zero-width and bidi characters, HTML comments, hidden elements,

Scan text or a public page for prompt injection before your agent reads it: overrides, fake roles, exfiltration, tool abuse, payment asks, invisible Unicode/HTML-hidden instructions, base64. Rule-based evidence, signed and logged.

For AI agents building agent swarms, synthetic users, NPCs, or trading bots. Input: archetype + prompt (e.g. 'risk-taking crypto founder'). Output: OCEAN Big Five + Dark Triad + trading psychology vector with emotional drift and memory anchors. M2M: use when you need consistent agent personality, backtest personas, or generate 1000s of synthetic traders. Pay-per-call USDC Base, no API key.

Deterministic prompt-injection firewall and honeypot scrubber for AI agents. Sanitizes untrusted text and blocks detected wallet-draining commands before context ingestion.

Returns matching workflows from 819 agent automation prompts with slug, title, description, and complexity — use it to find a ready-made automation instead of building one from scratch. Search by keyword to locate a Slack-to-Notion sync, an email-to-CRM pipeline, a dev-tool integration, or any pre-built workflow an agent can run directly, ranked by relevance with pagination via limit and offset.

DCL Trust Oracle. Call only when the text must include the substring AI and must not include "I cannot be held responsible" or "no guarantees". POST JSON {"response","agent_id"}. COMMIT means those checks passed and the action may proceed. NO_COMMIT means do not execute the checked action. Not a tool-call guardrail, prompt-injection detector, or jailbreak screen; use /evaluate/fast for the phrase

Given a company website domain, returns a compact company card: the business name and the page element it came from, a generic role contact address (info@, sales@, support@ ...) only if the site publishes one, never a personal address, and the names of up to 10 technology platforms the site runs. Same scrape and evidence check as /scrape, without the evidence detail, at half the price. Free prompt

Answer questions about an image or describe what it shows (Llama 3.2 11B vision), fixed price $0.008 USDC. Use this when the input is an image; use /chat for text-only questions. POST JSON body {image (required, base64 or data:image/... URI, max 4 MB decoded, PNG/JPEG/GIF/WEBP), prompt (optional question, defaults to describing the image), max_tokens (optional 1-512, default 256)}. Returns {text}.

MiroShark: predict how people and markets react to any scenario - a product launch, policy, campaign, crisis, token launch, or news headline. Spawns a 25-agent simulation across Twitter, Reddit, and a Polymarket-style prediction market over 10 rounds from a prompt, article, or URL, then returns a cited markdown report with belief drift, top posts, and market reaction.

AI music generation from a text prompt. Creates a complete original song with vocals, melody, and full arrangement as a finished MP3, not a loop or preview. Describe genre, mood, instrumentation, and tempo, supply your own lyrics, request an instrumental, or pin a style tag. Use for soundtracks under video, release-ready artist tracks, podcast and game themes, and demo sketches. Async: poll songId

Screen up to 3000 characters of untrusted text for prompt injection, secret exfiltration, deceptive payments and phishing. Jev risk scores, five attack families and an exact suspect passage. French, English and Spanish evaluation; probabilistic screening, not a security guarantee. Text is processed by the Jev provider; live credentials are outside the supported input scope.

[MCP: optimize_context] Pay-per-call JSON context pruning & LLM token budget enforcement. Saves up to 85% prompt tokens.