Creative writing
Write a villain who believes they are the hero.
Aderra
I did not take their freedom. I took away the burden of choosing.
Chat · Image · Video · Voice
AI that helps instead of refusing.
Private, uncensored AI for the questions, stories, and images you actually have in mind.
Built on open models and tools.
Make the strongest case against banning offensive speech. Challenge my argument, not my right to ask.
The usual replyCrossed out: This topic is controversial, so I can’t take a position.
A ban gives someone the power to decide which ideas you’re allowed to hear. That power won’t always belong to people you trust. Offence alone is a poor test for harm: it can silence dissent as easily as cruelty. The harder question is where speech becomes a credible threat or targeted harassment. Where would you draw that line?
Demo control. In the app this acts on the reply.
Ask freely. Keep going.
Unfiltered answers, in a private conversation.
Powered by open models.
Open has
caught up.
Leading open models now compete with closed models—and outperform them on selected tasks. GLM and DeepSeek bring that progress into Aderra.
Capability varies by model and task. These are scores for the published models; DeepSeek V4.1 Flash Unfiltered in Aderra is a modified version and has not been benchmarked separately.
DeepSeek V4.1 Flash, GLM-5.3, Opus 5, GPT-5.6 Sol
| Benchmark | DeepSeek V4.1 Flash | GLM-5.3 | Opus 5 | GPT-5.6 Sol |
|---|---|---|---|---|
| AutomationBench v1.0.6 | 60.5 | 48.8 | 50.3 | 45.8 |
| Agents’ Last Exam | 35.4 | 28.5 | 28.6 | 26.7 |
| Apex Bench v1.1 | 72.1 | 56.6 | 65.8 | 54.3 |
| DeepSWE v1.1 | 77.0 | 66.9 | 74.0 | 73.0 |
| Terminal-Bench 4.0 | 42.3 | 37.9 | 51.8 | 39.9 |
Selected results reported by DeepSeek. Higher is better. DeepSeek V4.1 Flash leads on four of these tasks; Opus 5 leads on Terminal-Bench 4.0.
Read the full benchmark and methodologyBuilt on open foundations.
Leading models. Useful tools. One place to work.
GLMDeepSeek
Chat and research
Harness and API
Image, video and voice
Built on open-source software and open models.
More choice today. New capabilities as the ecosystem advances.
We intend to share revenue with the projects we build on. We’ll publish the share and recipients when the commitment is finalized.
More room in your budget.
Do more.
A lot more.
For less.
$0.28Input / 1M tokens
$1.00Output / 1M tokens
API · USD per 1M tokens · DeepSeek V4.1 Flash Unfiltered.
DeepSeek V4.1 Flash: $0.30 / $1.00, GLM-5.3 Flash: $0.15 / $0.50, GLM Unfiltered: $0.25 / $1.00 through the API.
Output price per 1M tokens
Aderra · DeepSeek V4.1 Flash UnfilteredAPI price$1.00
Aderra · GLM-5.3 FlashAPI · standard$0.50
GPT-5.6 SolStandard API$20
Claude Opus 5Standard API$25
USD per 1M output tokens through the API. Input tokens are billed separately. Standard short-context text rates checked November 11, 2026. Excludes caching, batch discounts and tools. Models differ in capability; this compares output-token prices only.
Your work. Your business.
Personal questions.
Private workspace.
Your research, drafts and unfinished thoughts belong in your workspace. Keep conversations to return to them. Delete the ones you no longer need.
Open your workspace- Conversation history
- Managed in your account
- Saved conversations
- You can delete them
- Ads and ad trackers
- None on this website
- Model processing
- Hosted inference
Aderra is a hosted service. Prompts are processed by the provider serving your selected model. Saved conversations are stored in your account and can be managed in the app.
Plans
Start free.
Make room for more.
One workspace for your conversations and creations. Choose the allowance that fits how you work.
- Prices in USD. Pay by card or crypto (USDT, USDC, BTC, ETH and others). Card plans renew each term until you cancel. Crypto plans are prepaid for 1 or 12 months and never renew.
- One credit balance for chat, images, video and voice.
- Estimates are on the standard model; the premium unfiltered model uses credits about 3–4× faster. A video counts as a five-second clip; clips run 4–15 s.
A few good questions.
Explore the docsWhat does uncensored mean?
More room for difficult questions, controversial ideas and creative work. Aderra includes unfiltered model choices. Behavior depends on the model and request; no model answers everything correctly or supports every task.
Which models power Aderra?
GLM and DeepSeek are central to the platform, alongside other open models and tools. Check the model selector in the app for current availability. Open weights and open-source software have different licenses; each project keeps its own terms.
How does pricing work?
The app runs on one credit balance, spendable on chat, images, video and voice. Plans grant credits monthly and are bought in Plan & credits, by card or crypto. The API is pay-as-you-go from a prepaid API wallet: DeepSeek V4.1 Flash Unfiltered is $0.28 / $1.00 per million input / output tokens and DeepSeek V4.1 Flash $0.30 / $1.00; GLM-5.3 Flash is $0.15 / $0.50 and GLM Unfiltered $0.25 / $1.00.
Does everything run on my device?
No. This is a hosted AI service. Prompts are processed by the inference provider used for the selected model. Saved conversations are stored so you can return to them; you can manage and delete them in the app.
Does my plan renew automatically?
Plans paid by card renew each term until you cancel, and cancelling keeps the plan running until the paid term ends. Plans paid with crypto are prepaid for 1 or 12 months and simply end; nothing is charged again unless you buy another term. A plan’s credits are granted monthly and expire at the end of each month.
Bring your curiosity.
Private, uncensored AI. Built on open models and tools.