Skip to main content

Ideon

AI content writer for multi-channel publishing

Ideon turns one idea into multiple publish-ready outputs with consistent style, optional visuals, and research-enriched links.

It is designed for teams who need to ship high-quality content across channels without manually rewriting everything for each format.

Why Ideon​

  • Write many formats from one idea: article, blog post, newsletter, Reddit post, LinkedIn post, X thread, X post, and landing-page copy.
  • Keep voice consistent with style controls across all outputs.
  • Add depth with planning briefs, link enrichment, and generated images for article-led runs.
  • Iterate quickly with resumable runs, job files, and local preview.

Quickstart​

npm i -g @telepat/ideon
ideon write "Why async Rust is worth learning" --primary article=1 --secondary x-thread=2 --secondary x-post=1 --style technical

See Installation and Quickstart for full setup including credentials.

What Ideon produces​

  • One generation directory per run (timestamp + slug)
  • One or more markdown outputs (article-1.md, x-thread-1.md, x-post-1.md, etc.)
  • job.json run definition metadata for reproducibility
  • generation.analytics.json with stage and run metrics
  • Shared image assets rendered via a Replicate T2I model for the primary output (article primary includes inline images)

Documentation​

SectionWhat's inside
Getting StartedInstall, configure credentials, run first multi-output generation
GuidesConfiguration, writing framework, job files, pipeline stages, output structure
ReferenceCLI flags, content targets, styles, environment variables, supported T2I models
TechnicalArchitecture, LLM/image pipeline, testing
ContributingDev setup, adding models/stages/settings, releasing

Required credentials​

  • OpenRouter API key — for LLM calls (planning, article writing, channel outputs)
  • Replicate API token — for image rendering

Store them interactively: ideon settings — or set IDEON_OPENROUTER_API_KEY / IDEON_REPLICATE_API_TOKEN as environment variables.

Use --dry-run to test pipeline orchestration without making API calls.