A little less prompting. A lot more doing.

Big ideas.
Meet your
next move.

Your favorite AI models. A workspace that gets to work. Chat, research, write code, or give an agent a mission—and bring the results home.

One-time plans from $9.00. Or bring your own Anthropic key.

Room for your favorite mindsAnthropicOpenAI
THOUGHT ACTION SOMETHING REALStep inside

One place.
Every next step.

Keep the conversation, the tools, and the finished files together. Switch between supported Anthropic and OpenAI models without starting your workspace over.

Explore an illustrative Orrerie workspace preview
Interactive example
A new ideaChoose your model
You

Help me turn these notes into a launch plan.

project-notes.pdf
Let's give that idea a starting line.

We can shape the story, map the milestones, and build a plan you can share.

01 Define the audience & the promise02 Turn the ideas into a clear sequence03 Make room for feedback
Where should we take this next?
Agent modeA goal, with a trail
EXAMPLE MISSION

From rough notes
to a launch kit.

Give it a goal. Review the steps. Collect the files.

  1. Read the supplied project notes
  2. Shape a plan and draft the materials
  3. Save the work for your review
Missions run within approval, tool and time limits.
WorkspaceExample deliverables
MDlaunch-plan.mdThe plan, ready to review
CSVmilestones.csvA little structure for the big picture
TXTresearch-notes.txtYour references, kept together

In the console, download individual files or the whole workspace.

Your models, in reach.

Supported Anthropic and OpenAI models, one searchable catalog.

Know what the work costs.

See metered AI usage, context and cost beside your conversations.

Take the work with you.

Download generated files and manage your conversation archive.

See the actual workspace in motion.Orrerie product tour · 88 seconds

Bring the idea.
We brought the tools.

Documents. Research. Code. The things you need to get from a question to work you can use, all in one workspace.

Reference image
Project brief.pdfPDF
Every idea starts
with something.
Example attachments

More than a text box.

Bring images, screenshots, PDFs or a long text paste. Ask compatible models to read, explain and work with your material.

Follow the source.

Research authorized sources. Agent’s guarded browser can open permitted pages, fill forms and download files.

analysis.pyExample code
import pandas as pd

data = pd.read_csv("source.csv")
summary = data.describe()
summary.to_csv("summary.csv")
From source material to a real file.

Think it. Build it. Run it.

Use permitted code to process data and create files. Work with Python, compatible packages and ffmpeg for audio and video tasks inside the agent sandbox.

The first ideaConversation
A different direction
The version worth keeping
Example conversation

Good work has a history.

Search your private archive, edit a message and re-run from there. Pin what matters or set it aside for later.

Input & outputToken usage
ConversationContext & cost
Your planAllowance & access

Know what goes into it.

See token usage, context and cost alongside your work. Review your remaining allowance and access period in Usage & billing.

Your services. Your permission.

Connect supported services for authorized tasks. Saved secrets are encrypted and masked on readback. Consequential actions may require your approval.

Tool availability depends on the selected model, plan, connection and provider. Explore the product

A question gets
an answer.
A mission
gets to work.

Research authorized sources. Run permitted code. Build real files. Give a bounded agent mission a clear outcome, and follow the work as it happens.

Create an account Included with Orbit and Supernova
Example mission Agent
You

Turn this CSV into a clear report with charts and notes I can share.

source.csv
Orrerie Agent

A clear outcome. A workspace for the work.

  1. Read the source materialInspect the attached file and its structure.
  2. Work through the analysisRun permitted code and review the output.
  3. Build the deliverablesPrepare files for your review.
Example deliverables Files
PDFreport.pdfA report to review
CSVsummary.csvThe analysis, as a file

Illustrative workflow and filenames. Actual steps and results depend on the task.

Close the tab.
Keep the mission.

A launched mission runs on the server. Return to reconnect, inspect its progress or collect completed work. Closing the tab does not itself cancel it.

A workspace
with boundaries.

Each conversation has a scoped workspace. Packages, browsing and connected services remain subject to sandbox restrictions and authorization.

The work is yours
to download.

Collect reports, PDFs, spreadsheets, charts or code. Download individual files, the workspace as a ZIP and available mission trails. Sensitive details are redacted.

What can stop a mission?

Missions can stop for time or capacity limits, errors, approvals, safety restrictions or provider availability. Completion is not guaranteed. Review important AI output before using it, and only automate sites, systems and data you are authorized to access. Read the acceptable use policy.

The whole clip.
The whole story.

Ask what happened. Video Watch follows visual changes, reads on-screen text and keeps repeated moments in their proper order.

Follows the visual sequence Reads visible text Preserves separated repeats

Visual review does not analyze speech in the audio. Important AI output should be checked.

See Video Watch in action

From “what if”
to “let’s go.”

No local installation is required for ordinary hosted use.

  1. Sign in

    Use Google or an email account with a verified address. Optional two-factor authentication adds an extra account check.

  2. Choose access

    Buy a one-time plan through OxaPay crypto checkout, or connect your own Anthropic API key for compatible chat.

  3. Pick your model

    Search and filter the supported catalog. Select model-specific reasoning controls where available.

  4. Work or delegate

    Chat with your files and selected tools, or use Agent mode on an eligible plan. Review its progress and collect the work.

Different minds.
Same momentum.

Find supported Anthropic and OpenAI models in one searchable catalog. OpenAI models work in normal chat and can also power Agent mode.

Open the model catalog
AnthropicThoughtful by design.OpenAIIdeas into action.

Availability, in the console

The console shows the supported catalog and current upstream availability when a provider check is available. Model names, limits and rates can change.

Controls for the selected model

Thinking and effort settings appear for compatible models. Provider restrictions and account access determine which choices are available for a task.

Plan or your own Anthropic key

Plans include a metered AI-processing allowance. Own-key chat has no Orrerie plan fee; Anthropic bills that API usage separately.

One-time plans.
No automatic renewals.

Paid plans include supported Anthropic and OpenAI chat models. Agent mode is included in the plans marked below.

Comet

Casual chat and quick answers across supported models.

$9.00/ 30 days

$3.50 AI-processing allowance

  • Supported Anthropic + OpenAI models
  • Images, documents & compatible tools
  • Visible metered usage
  • Everyday AI chat
Get started One purchase. No auto-renewal.

Supernova

For power users who live in the console all day.

$49.00/ 30 days

$22.00 AI-processing allowance

  • Supported Anthropic + OpenAI models
  • Images, documents & compatible tools
  • Visible metered usage
  • Bounded Agent missions
Get started One purchase. No auto-renewal.

Bring your own Anthropic key

No Orrerie plan fee for compatible chat. Anthropic bills your API usage directly.

Current checkout

Supported cryptocurrencies through OxaPay. Card payment is not currently available.

Know what you are buying

The allowance is an internal service limit, not money or transferable credit. It expires with the access period.

View full pricing & payment details

Fair questions.

More about access, plans and how your work is handled.

Contact support
Do I need my own API key?

No. Plans include access to supported AI models. You can instead connect your own Anthropic API key for compatible chat. It is validated and encrypted in account-scoped server storage; Anthropic bills that usage separately.

How do payments work?

Current paid checkout uses supported cryptocurrencies through OxaPay. Choose the exact asset and network shown. Activation follows sufficient provider and network confirmation; card payment is not currently available. See pricing and the refund policy.

Do plans auto-renew?

No. Plans are one-time purchases. A purchase during an active period adds its allowance and access days. The allowance is an internal service limit and expires with the resulting access period. Check Usage & billing for your remaining access.

Which plans include Agent mode?

The plan cards show each plan’s access. Comet includes chat; Orbit and Supernova also include bounded Agent missions. Tool and model availability depends on compatibility, provider availability and applicable limits.

Can a mission continue when I close the tab?

Yes. A launched mission runs on the server, so closing the tab does not itself cancel it. You can return to reconnect and inspect the outcome. Time, capacity, approvals, errors, safety and provider limits can still stop the mission.

Where is my data stored?

Private records and workspaces are scoped to your account and conversation, outside the public web root. Provider processing, retention, downloads and deletion are described in the Privacy Policy. See Security for implemented controls.

Which services receive my content?

Orrerie sends the content needed for a task to the selected model provider and to services or websites you deliberately connect or invoke. The Privacy Policy describes provider processing and the applicable data controls.

How does sign-in work?

Use Google or email and password with a verified email address. Accounts support optional two-factor authentication with an authenticator app and backup codes, plus self-service password reset.

What rules apply to browser automation?

Use it only on sites, systems, accounts and data you own or are authorized to access. Unauthorized scraping, access-control bypass, credential abuse, spam, fraud and unsupervised financial transactions are prohibited by the Acceptable Use Policy.