Free
- Concurrent pipelines
- 1
- Parallel inputs
- 1
- Results per month
- 10
- Tests per day
- 3
AI CONTENT PIPELINE FOR WEB PUBLISHING
Sources, RAG, AI steps, review and publishing in one visible workflow.
Execution
The free plan includes one slot and one thread. A template, provider connection and WordPress or CMS endpoint are enough for a complete test run.
plans
| Features | Free Free | Top · best choice Starter 10 USDT / period | Pro 29 USDT / period | Agency 79 USDT / period |
|---|---|---|---|---|
| Concurrent pipelinesPipelines that can run at once | 1 | 2 | 5 | 10 |
| Parallel inputs per pipelineInputs processed in parallel | 1 | 2 | 4 | 8 |
| Successful results per monthFinal production results | 10/month | Unlimited | Unlimited | Unlimited |
| Template tests per dayTesting before production | 3/day | Unlimited | Unlimited | Unlimited |
| Get started | Get started | Get started | Get started |
how it works
platform layers
How the platform is wired: five layers from raw input to a published page. This is the anatomy, not the pitch — what each layer takes in and returns.
A run can begin with keywords, search results, exact URLs, news, media or output from an earlier step.
Deterministic steps clean, split, combine, deduplicate and reshape data before a model sees it.
Every step has its own provider, model, prompt, temperature, token limit and output key.
Prompts, domain rules, variables, tests, debug output and queue state remain available for inspection.
The final step sends structured fields to WordPress, the AGD CMS client or another compatible endpoint.
pipeline example
The exact steps depend on the template. A research-backed SEO workflow can follow this sequence.
templates
Importable examples cover simple generation, source-backed research, multilingual output and structured product posts. Every step remains editable after import.
comparison
An AI article generator returns text. A content pipeline also handles the stages before and after writing.
| AI article generator | Content pipeline |
|---|---|
| One prompt | Separate research, writing, review and publishing steps |
| One model | Provider and model selected per step |
| Prompt-only context | Search, source pages and RAG context |
| Text response | Structured title, description, content and extra fields |
| Hidden processing | Visible inputs, outputs and debug data |
| Manual copy and paste | WordPress and CMS publishing through API |
| Single generation | Templates, keyword sets, queues and schedules |
| Provider token bundle | Own API keys without token markup |
AGD Flow helps publishers, SEO operators, affiliate teams and developers build an AI content pipeline that turns keywords into RAG-ready SEO data, generates source-backed articles with OpenAI, Gemini, Anthropic, DeepSeek or Qwen models, and auto-publishes structured content to WordPress. The free plan includes one slot and one thread — no card required.
faq
Short answers about sources, RAG, models, costs and WordPress publishing.
It is a repeatable workflow that turns an input such as a keyword or URL into structured content. The workflow can collect sources, prepare RAG context, run AI and review steps, map fields and publish the result.
Start from a ready template: connect your AI provider keys, set the keyword or URL input, adjust prompts and review rules, then run the queue. The pipeline collects search results and sources, prepares RAG context, generates the article step by step and publishes it to WordPress or another CMS.
A generator mainly returns text from a prompt. A pipeline also controls source collection, processing, model choice, review, queues and delivery to WordPress or another CMS.
RAG supplies retrieved source material before writing. It does not guarantee accuracy, but it makes the context visible and reduces dependence on unsupported model memory.
It is source material prepared for retrieval-augmented generation: search results, fetched pages, cleaned fragments and limited-size context blocks. The pipeline collects and formats this data so the writing model works from actual sources instead of memory alone.
Yes. Each AI step can use its own provider and model. Planning, writing, review and JSON formatting do not need the same model.
Article Form maps title, description, content, image, tags, categories and custom fields. The mapped data is sent to WordPress through API or the AGD CMS client.
Yes. The workflow itself runs in AGD Flow — research, RAG context, AI steps and review — while WordPress stays the publishing destination. Mapped fields are pushed through the WordPress REST API or the AGD CMS bridge, so posts, categories, tags and custom fields appear without manual copy-paste.
Yes. Keyword sets, queues, schedules and publication limits control repeated runs. Each run keeps its own sources, context, status and publishing response.
Cheaper models can handle planning and extraction while stronger models handle drafting or review. Supported OpenAI and Anthropic prompt caching can reduce repeated input cost by up to 90 percent, and supported Batch APIs can reduce non-urgent model cost by up to 50 percent.
No. Providers are connected with the account owner's API keys, and token charges remain with those providers.
articles and guides
Practical material on AI content pipelines, RAG for SEO, WordPress workflows and controlled bulk publishing.
A practical explanation of AI content pipelines: inputs, source collection, RAG, model steps, review and structured publishing.
Read →A practical guide to RAG for SEO: source selection, cleanup, chunking, freshness, review and a visible content workflow.
Read →A practical WordPress AI workflow for sources, RAG, review, field mapping, drafts, custom fields and controlled auto-publishing.
Read →A practical workflow for bulk AI articles: keyword queues, RAG, duplicate control, review, cost limits and WordPress auto-publishing.
Read →Make and n8n are general automation platforms. AGD Flow is narrower and deeper for source collection, RAG, AI steps, Article Form and WordPress publishing.
Read →When Jasper as an AI content writer is not enough: source collection, RAG context, model-per-step and WordPress auto publish in a content pipeline.
Read →When to move from Agility Writer to a content pipeline with source collection, RAG context, model-per-step and WordPress auto publish.
Read →Practical comparison of Byword, Koala and Cuppa as AI SEO writers, and where a pipeline like AGD Flow fits when sources, RAG and WordPress publishing matter.
Read →