Now in Early Access

The agent-first workspace for reproducible research

From intake to publish—capture decisions, artifacts, and provenance by default. Research progresses through defined stages, with AI agents assisting and humans governing.

No credit card required • Free for individual researchers

Research with integrity, by design

Not another AI tool that makes claims. A structured workspace that captures the decisions, artifacts, and reasoning behind your research.

Stage Machine

12 defined stages from Intake to Publish. Research progresses visibly, with checkpoints at each transition.

Agent Fleet

6 specialized AI agents—Planner, Literature, Analysis, Manuscript, Reviewer, QC—working together, human-governed.

Provenance

Every artifact hashed and tracked. Know exactly what changed, when, and why. Full audit trail by default.

Lanes

Separate exploratory from confirmatory work. Pre-registered protocols stay locked. Discoveries stay labeled.

Security

Content firewall protects against injection. Data classification routes sensitive work to appropriate models.

12 stages from intake to publish

Research progresses through defined stages. Each transition captures a checkpoint. You always know where you are and what's been done.

S0
Intake
S1
Plan
S2
Literature
S3
Synthesis
S4
Protocol
S5
Data
S6
QC
S7
Analysis
S8
Figures
S9
Manuscript
S10
Review
S11
Publish

S7: Analysis

Statistical analysis execution

Learn more about the workflow

Six specialized agents, human-governed

AI handles routine work. You make the judgment calls. Every agent action is logged, and sensitive operations require your approval.

Planner

Breaks missions into tasks

Literature

Searches papers, extracts claims

Analysis

Runs statistical methods

Manuscript

Generates draft text

Reviewer

Critically evaluates work

QC

Validates data quality

Meet the agents

Built for accountability, not magic

SciAI helps you capture and verify your research process. It does not guarantee correctness—that's your job. We make it auditable.

Hashed Artifacts

Every output is SHA-256 hashed for integrity verification

Lockable

Lock artifacts to prevent modification

Approval Gates

Sensitive changes require human approval

Audit Trail

Full history of runs, approvals, and stage checkpoints

View our security model

A research project, halfway through

Explore our demo project: a Cancer Immunotherapy Meta-Analysis with papers, missions, and agents at work.

Watch the 2-minute tour

Book a Live Demo

Simple, transparent pricing

Start free during beta. Scale when you're ready.

Individual

For solo researchers

Free
during beta
  • Core workflow (S0→S11)
  • 1 active project
  • Agent Fleet access
  • Basic provenance
  • Community support
Get Started

Enterprise

For institutions

Custom
annual contract
  • Everything in Lab
  • SSO / SAML
  • Compliance features
  • Self-host option
  • Dedicated success manager
  • SLA guarantee
Get Started

Common questions

Chat-with-PDF tools let you ask questions about documents. SciAI structures your entire research workflow—from literature search to manuscript—with provenance, stage tracking, and human governance at every step. It's not about chatting; it's about building an auditable research process.
Our content firewall sanitizes all external content before LLM processing. We detect and redact known injection patterns, and we use domain allowlists for web fetching. Sensitive data can be routed to local models only.
Data classification determines routing. PUBLIC and INTERNAL data may use cloud providers (Gemini, OpenAI). CONFIDENTIAL uses only trusted providers with enterprise agreements. RESTRICTED routes exclusively to local models.
No. SciAI ensures process integrity and traceability—it captures what you did, when, and why. Scientific correctness remains your responsibility. We help you prove your process was rigorous, not that your conclusions are right.

Ready to make your research auditable?

Book a demo and see how SciAI can structure your workflow.

Book a Demo