🧪 This platform is in early beta. Features may change and you might encounter bugs. We appreciate your patience!
Everything you need to know about TellDear — from getting started to the full aspect taxonomy.
TellDear is a media and argument analysis platform. It helps you see through rhetoric, detect flawed reasoning, and train your critical thinking skills — powered by a taxonomy of 535 aspects across 6 dimensions.
You can use TellDear without an account for browsing the taxonomy and reading articles. AI-powered apps work immediately in free mode — or register and add your own API key for the best results.
TellDear offers 25 AI-powered apps organized into categories:
💡 Most apps require an API key. The heuristic mode in the Text Analyzer works without one.
TellDear's taxonomy contains 535 aspects organized in 6 dimensions:
Each aspect has binary verification steps, a formal logic pattern, examples, and related aspects. Browse them all in the Aspect Directory.
TellDear supports multiple AI providers. You can use keys from Anthropic, OpenAI, Google, Mistral, xAI, DeepSeek, Groq, or OpenRouter.
Your key is encrypted and used only for your own requests. We recommend setting a monthly spending limit (e.g. $5–10) at your provider.
You can add multiple keys — if one fails (rate limit, quota), TellDear automatically tries the next one.
💡 Our recommendation: Start with Google Gemini (free tier, great quality) or Anthropic Claude (best analysis quality). Most users spend less than $5/month.
sk-ant-)Cost: ~$3/1M input tokens (Sonnet). Set a $5–10/month spending limit in Settings → Limits.
Tip: Best overall quality for critical thinking analysis. Claude excels at nuanced reasoning.
AIza)Cost: Free tier: 15 requests/min for Gemini 2.0 Flash. Paid: ~$0.10/1M input tokens (Flash) — extremely affordable.
Tip: Best price-performance ratio. Generous free tier — great for getting started without spending anything.
sk-)Cost: ~$2.50/1M input tokens (GPT-4o). Set a monthly limit under Settings → Billing → Usage limits.
Tip: Great all-rounder. Strong at structured analysis and multilingual content.
sk-)Cost: ~$0.27/1M input tokens (V3) — one of the cheapest frontier models. Cache hits even cheaper at $0.07/1M.
Tip: Excellent value. DeepSeek R1 for reasoning, V3 for general analysis. Very affordable.
Cost: Free tier available for experimentation. Paid: ~$0.10/1M tokens (Small) to $2/1M (Large).
Tip: European provider (Paris). Free tier for prototyping. Strong multilingual support.
xai-)Cost: ~$3/1M input tokens (Grok 3). $25/month free credit for new accounts.
Tip: Strong reasoning model. Good for complex argument analysis.
sk-or-)Cost: Free models available (rate-limited). Paid: pass-through pricing from each provider + small margin.
Tip: Access 300+ models with one key. Great if you want to try different models without separate accounts.
gsk_)Cost: Free: 1,000 requests/day. Ultra-fast inference on Llama models.
Tip: Fastest inference available. Free tier is generous. Uses open-source models (Llama 3.3 70B).
Any OpenAI-compatible provider works. If your provider has an OpenAI-compatible API, it will likely work with TellDear.
✨ Browsing the taxonomy, the Aspect Directory, the Knowledge Graph, and all articles is completely free — no account or key needed.
Have questions? Check the FAQ for common questions about the platform, apps, privacy, and more.
For feedback or issues, use the Contact page.