AI newsroom
A newsroom where you control every step from source to publication.
Gastte doesn't sell you AI credits. You enter your key and pay the provider directly; in the dashboard you decide which model runs each task and how much is spent.
Workflow
News in seven steps.
- SourcesRSS feeds are fetched regularly; keyword filter per source.
- Incoming NewsCopies of the same story are clustered; age and relevance filters.
- RewritingA model per task; original, copyright-compliant stories in your publication's language; source attribution.
- VerificationClaims in the story are extracted and flagged; checked on the verification screen.
- Approval queueOptional: every story goes through editor approval before publishing.
- PublishThe pre-rendered page is refreshed; sitemap and IndexNow notification.
- DerivativesHeadline, summary, SEO, tags, cover image, social copy, podcast.
Safe automation
Don't trade accuracy for speed.
Copyright rule
The source text is never copied; it is rewritten originally and source attribution is added.
Approval queue
You can choose approved publishing instead of automatic; rules by category and source.
Verification
Claims are extracted and flagged; the verification screen makes checking easy.
Automation rules
Which source, which category, what time; automatic or approved — you decide.
Providers
10 providers, with your own key.
You can start with providers that have a free tier and switch to paid whenever you like. Model names change, so they aren't listed here; the dashboard shows the current list.
- Anthropic ClaudeÜcretli
- CerebrasÜcretsiz katman
- Cloudflare Workers AIÜcretsiz katman
- DeepSeekÜcretli
- Google GeminiÜcretli
- GitHub ModelsÜcretsiz katman
- GroqÜcretsiz katman
- NVIDIA NIMÜcretsiz katman
- OpenAIÜcretli
- OpenRouterÜcretsiz katman
| Provider | Pricing | Text | Image generation | Embeddings | Vision |
|---|---|---|---|---|---|
| Anthropic Claude | Paid | — | — | — | |
| Cerebras | Free tier | — | — | — | |
| Cloudflare Workers AI | Free tier | — | — | ||
| DeepSeek | Paid | — | — | — | |
| Google Gemini | Paid | ||||
| GitHub Models | Free tier | — | — | ||
| Groq | Free tier | — | — | — | |
| NVIDIA NIM | Free tier | — | — | ||
| OpenAI | Paid | ||||
| OpenRouter | Free tier | — | — | ||
| Source: dashboard provider profiles, 2026-10-27. | |||||
Cost control
Spending limits in the dashboard, no surprise bills.
Budgets and limits
Daily and monthly spending limits per provider; when a limit is reached, work stops and you get an alert.
Cost dashboard
Spending per task and provider; the job log records every call.
Backups
If the primary provider goes down, a backup provider takes over; you won't notice.
Prompt management
A ready-made catalog, your own edits.
Ready-made prompts come for every task; copy them and adapt them to your publication's voice. The prompt test and model comparison screen lets you see the result before publishing.