AxWiseFind the right questions
Propose a discovery scope, stakeholder roles, uncertainties and a linked interview guide. Revisit the scope as you learn.
prepare_discoveryA specialist for your AI workspace
Better questions.
Clearer evidence.
Stronger products.
AxWise brings structured product discovery to your agent. Turn a brief, selected research and interviews into personas, insights and a product plan you can build on.
Open-source extension. Local workflow. Your chosen model.
A simpler homeMeet Orqanix
The desktop experience is now in Orqanix.
AxWise is the focused discovery extension inside it. This site is home to the open-source project, examples and standalone installation—not a second web app.
A useful specialist, not another orchestrator
From an open question
to a working product brief.
Choose one capability, or ask your agent to connect the steps. Standard and deeper analysis are bounded options for the task you select.
Bring perspectives into the room
Generate synthetic persona cohorts, simulate interviews and talk with an exact saved persona about a selected document.
generate_personas · simulate_interviews · chat_with_personaMake sense of selected evidence
Analyze interview quotations, themes, patterns, stakeholder sentiment and insights. Synthesize selected market sources and surface gaps.
analyze_interviews · research_marketTurn learning into a plan
Create an evidence-linked PRD, preserve its history through revisions and propose a delivery brief with acceptance checks and open decisions.
create_prd · create_delivery_briefInside the loop
Your agent
stays in charge.
Goose, Codex or another compatible MCP host owns the conversation, tools and approvals. AxWise is called for a specific discovery deliverable—not for every message.
Explore the source- Your hostChat, context selection, tools and permissions
- AxWise extension · localWorkflow logic, validation, state and saved artifacts
- Configured model · remoteInference through your provider or the Orqanix gateway
“Local” describes where the workflow runs. Selected material is sent to the configured model; this is not an offline-inference claim.
Start with what you need
Small prompts.
Substantial work.
No special command language. Ask for the result you want in your normal chat. Your host chooses the tool and may ask you to clarify the scope.
Get a project off the ground
“We’re building a booking tool for independent studios. Propose the discovery scope, stakeholder groups and interview questions before we decide what to build.”
Understand what you heard
“Analyze these selected interviews. Show themes, patterns, stakeholder sentiment and conflicting needs, linked to supporting quotations.”
Explore another perspective
“Create three synthetic personas from this scope, simulate the interviews, then let me discuss this exact draft with the studio manager persona.”
Move toward implementation
“Use this saved analysis and selected market evidence to create a PRD. Add onboarding requirements without losing the earlier ones, then prepare an outsourcing brief.”
Simulations are labeled synthetic. AxWise keeps their provenance with the resulting artifacts. Model-generated interpretations still need your judgment.
One engine, two ways in
Use the desktop.
Or bring your own host.
Included in Orqanix
The desktop bundles AxWise and its runtime. Enable the extension when you want its specialist capabilities, and keep working in the same chat.
Orqanix adds its own sign-in, model access and shared Results experience. Its service terms and model access are separate from the open-source license.
Get OrqanixAxWise 0.3.0
Two launch paths, the same engine. These commands use the versioned GitHub release artifacts—not an npm or PyPI registry listing.
Launch with npx
npx --yes --package=https://github.com/AxWise-GmbH/axwise-flow/releases/download/axwise-extension-v0.3.0/axwise-extension-0.3.0.tgz axwise --config /absolute/path/axwise.jsonOr launch with uvx
uvx --from https://github.com/AxWise-GmbH/axwise-flow/releases/download/axwise-extension-v0.3.0/axwise_extension-0.3.0-py3-none-any.whl axwise --config /absolute/path/axwise.json- Both require Node.js 22+ and Python 3.11+ (uv can provide Python).
- Install uv before using either launch path.
- Provide your own model credentials in your local configuration; do not paste secrets into chat or public config files.
Configuration and host examples ↗
Verified download checksums (SHA-256)
npm archive · 157,987 bytes
a17caf788dd42e7979f281f2dce6aa3ec527ad6bd1b3a8c8a7f7b2d72c22fe20Python wheel · 152,383 bytes
d3c13786ce62de85a8e7dc6ef380c9cb0de61ca6b6477409133b579d36fe336fThe code is free to use and modify under Apache 2.0. Model-provider usage may incur charges. Standalone use does not require an Orqanix account.
Clear boundaries
What changes.
What stays yours.
Does AxWise take over my agent?
No. It exposes specialist tools through MCP, not a per-message router. The host decides when to invoke a tool and retains its own permission system. Ordinary chat, weather, news, browsing and coding stay with the host’s normal tools. Tool selection varies by host and model; it is not a perfect intent classifier.
Where do the research and results live?
Workflow state and immutable Markdown/JSON artifacts are saved locally. Selected inputs go to the configured model for inference. Revisions link back to their earlier result. Orqanix shows them in its Results panel; another host may display tool output and local files differently. There is no automatic cloud session sync in the extension.
Does AxWise search the web for me?
No. The market capability synthesizes exact selected sources, including their URLs and retrieval dates, and can propose questions or searches for missing evidence. Your host uses its own search tools and decides what to pass back. AxWise is not a replacement search engine.
Can I use it outside Orqanix?
The standalone extension speaks MCP over standard input/output for compatible hosts such as Goose and Codex. It uses the same core engine, with your own provider credentials. Host configuration, discovery, permissions and result rendering differ. Follow the repository’s setup examples for your host.
What happened to the old AxWise web platform?
The project is being focused on its discovery and research capabilities as a local extension. The old dashboards and hosted orchestration are not part of this package. Video/pre-call tools and regional dashboards are not advertised as extension features. Existing repository history remains available; this is a new direction, not a promise of full legacy-web parity.
Keep your workflow. Add a specialist.