One workspace for the whole research lifecycle
Sapien is the first workspace to unite agentic AI research, real-time collaborative LaTeX, Typst & Markdown writing, Python, R & Jupyter notebooks that run in the browser, and a knowledge graph of everything you read — all grounded in your sources and governed by your institution.
The idea
One workspace for the paper, the sources, and the thinking between them.
Today a single paper is scattered across a writing app, a shared doc, a reference manager, a search tool, a notebook, a chatbot, and a folder of PDFs — none of which share context. Sapien replaces that stack with one project where writing, sources, code, and AI finally work from the same place.
For researchers, labs, universities, and research organizations that need faster work without giving up control over hosting, models, or data.
What only Sapien does
Each of these exists somewhere. Researchers stitch them together across half a dozen apps that never share a source, a citation, or a thought. Sapien is the first to put all four in one project — on the same context and the same AI.
Point an agent at your library and it searches, reads, synthesises, drafts, and writes the analysis code — every step grounded in real sources, fully reviewable, and yours to approve.
Co-author a thesis in LaTeX, a preprint in Typst, and notes in Markdown without switching tools or breaking the cursor. Multiplayer editing, one document model.
Execute notebooks in isolated containers right beside your draft. Figures, tables, and outputs save back to the project and stay citable.
Sapien links papers, claims, datasets, notes, and outputs into a graph you can search and explore — surfacing the threads between sources that a folder of PDFs never could.
You hold the dial
Sapien is a complete research workspace before a single token is generated. Real-time collaborative editing, in-browser computation, and the knowledge graph all work on their own. Turn AI up when it earns its place, down when policy requires, or off entirely — nothing breaks.
The complete system
Around those four capabilities sits the connective tissue of a real workspace — so a literature review or a paper moves from first source to final draft without ever leaving the project.
Keep project files, collections, graph views, agent sessions, collaborators, shared documents, and layouts together so a literature review or paper stays organized over time.
Write collaboratively in Markdown, Typst, LaTeX, Python, R Markdown, and Jupyter, with shared context for the whole research space.
Find relevant passages across files, collections, spaces, reference integrations, and prior sessions.
Explore connections between entities, claims, topics, files, and source segments in graph and card views.
Use agents to draft, revise, critique, summarize, and extend writing with approvals, skills, file mentions, forks, and exportable records.
Read PDFs, Word documents, spreadsheets, text files, and extracted segments. Run Python, R Markdown, and notebooks in isolated containers with saved artifacts.
Three working modes
Markdown, Typst, LaTeX, Python, R Markdown, and Jupyter live in one surface — so drafting, discovery, and executable analysis share the same sources, citations, and AI instead of living in separate apps.
Format preview
Swap between prose, typeset manuscripts, scripts, and notebooks without separating the work from the research space.
Draft in academic formats while files, notes, citations, and project context remain close to the manuscript.
Search across papers, notes, references, sessions, and extracted source segments without leaving the project.
Use writing agents, run code, and route selected tasks to approved model providers or local endpoints.
Agentic work
Hand the agent a bounded task and the right context, and it does the work: summarise sources, pressure-test claims, revise prose, prepare analysis code, and return artifacts you can review, approve, and trace back to their evidence.
Choose the files, notes, source segments, draft section, or notebook cells the agent can use.
Ask for a grounded synthesis of papers, interviews, policies, or source collections.
Generate structured revisions, flag weak claims, or propose a tighter argument.
Create Python, R Markdown, or notebook cells for analysis tasks and checks.
Keep humans accountable with saved records, exports, approvals, and follow-up tasks.
Deployment flexibility
Use public cloud, private cloud, on-premises infrastructure, or bare metal. Pair that with approved model providers, local endpoints, and storage patterns that match your policies.
Configure model providers, storage, and deployment patterns around your institution's policies instead of forcing every researcher into the same hosted stack.
Deployment
One platform, three ways to run it: self-host for full control, use the hosted workspace with zero setup, or deploy with enterprise support and governance for your whole institution.
Run Sapien in your own environment with control over infrastructure, data, models, and connected tools.
A hosted Sapien workspace for researchers who do not want to manage infrastructure.
Deploy Sapien with AcademicID support for organizations that need governance, rollout planning, and account controls.
Compare options
Each option is designed for a different level of control: who operates the stack, where data lives, and how much governance your organization needs.
| Capability | Self-hosted | academicid.io | Enterprise |
|---|---|---|---|
| Status | Coming soonFor teams that want to operate Sapien themselves. | Coming soonHosted by AcademicID for researchers and teams. | Coming soonSupported deployments for institutions and research organizations. |
| Best for | Technical researchers, labs, and teams that want full infrastructure control. | Researchers and teams that want Sapien without managing infrastructure. | Universities, libraries, labs, and organizations with governance requirements. |
| Operations | You deploy, update, monitor, and back up the system. | AcademicID operates the hosted service. | AcademicID supports deployment planning, onboarding, operations, and rollout. |
| Data and storage | Your infrastructure, databases, vector index, graph database, and storage policies. | AcademicID-managed hosted storage and service configuration. | Institution-aligned storage, data locality, retention, and security configuration. |
| Models and AI providers | Bring approved cloud, direct, or local model endpoints, and configure MCP servers for agent tool access. | Use hosted defaults and supported provider options. | Configure provider access around institutional model governance and procurement. |
| Admin and account control | Local admin control for your own deployment. | Standard account controls for individual researchers and teams. | Advanced admin and account controls over usage.Manage users, workspaces, permissions, usage visibility, and organizational policies. |
| Support level | Community documentation and release notes. | Hosted service support. | Deployment, onboarding, configuration, and operational support for organizations. |
| How to start | Join release updates for self-hosted access. | Join the waitlist for academicid.io. | Contact AcademicID to discuss institutional requirements. |
For academic organizations
Research AI needs sources, permissions, and a record of the work.
Sapien is for settings where data locality, model governance, source transparency, reproducible analysis, and auditability matter as much as convenience.
Deploy Sapien on the infrastructure your organization already trusts: public cloud, private cloud, on-premises clusters, or bare metal. Keep research workflows aligned with your security, storage, and governance requirements.
Join the waitlist
Get updates on hosted access, self-hosted releases, and enterprise-supported deployments for universities, labs, libraries, and research organizations.
Waitlist
Join the list for hosted access, self-hosted releases, and enterprise pilots.
Cited from your Zotero library
