Read beyond the abstract
Connect a paper’s claims with its methods, evaluation setup, and limitations. Keep the context that makes a finding useful.
Paper → Method → Limitation
AI RESEARCH COPILOT
Turn ideas into
evidence-backed
research plans.
Our focus is an AI-native research workspace connecting papers, hypotheses, and experiment design — helping researchers see what the evidence supports and decide what to test next.
01 / THE PRODUCT DIRECTION
A copilot for the work
between reading and doing.
Our product direction connects literature review and promising ideas with concrete experiments, bringing the evidence and the next research decision into one place.
Make evidence inspectable
Link important claims to sources and relevant passages. Surface conflicting findings and separate evidence from assumptions.
Claim ↔ Source
Plan the next experiment
Shape a broad idea into a testable hypothesis, with datasets, baselines, controls, evaluation criteria, and failure conditions.
Question → Test → Decision
Carry the reasoning forward
Keep the evidence and decisions behind each plan, so feedback and new results can inform the next iteration.
Evidence → Review → Revision
Share the research workflow you want to improve.
02 / FROM A QUESTION TO A PLAN
From a question
to a research plan.
Follow one example from a research question to a testable plan. Select a stage to explore the workflow.
RESEARCH WORKFLOW EXAMPLEEVIDENCETRELLIS / RESEARCH WORKSPACE
RESEARCH QUESTION
Can retrieval-grounded assistants produce more reproducible experiment plans?
03 / OUR APPROACH
Useful AI needs
something to stand on.
Evidence, uncertainty, and researcher judgment guide our approach to research tools. The reasoning behind a plan should be easy to examine.
Sources stay visible.
A useful research plan should let you inspect where its claims come from.
Uncertainty stays explicit.
Open questions, missing evidence, and failure conditions belong in the plan.
Researchers make the call.
The copilot should support human judgment, with review before consequential steps.
04 / WHAT COMES NEXT
Build the foundation.
Then grow the loop.
Our roadmap moves from grounded reading to research planning and connections with research tools. Each stage informs the next.
01First focus
Grounded reading
Start with paper organization and claim-to-source links. Build a small, reviewable literature-to-plan prototype.
02Next milestone
Research planning
Connect evidence to hypotheses and experiment designs. Refine the workflow through feedback from researchers.
03Longer-term direction
Tool-assisted iteration
Connect researcher-approved plans to data and research tools. Use observed results to guide revisions.
Researchers’ feedback guides our priorities and the connections we build.
05 / PROPOSED PLANS
Start with a question.
Choose room to grow.
PROPOSED MONTHLY PRICING
A starting point for how individual researchers and teams could use EvidenceTrellis. Tell us which plan would fit your workflow.
Explorer
$0/ month
Get a feel for a paper-to-plan workflow.
- Organize papers around a research question
- Inspect claims and their source context
- Draft a focused question and next-step outline
Researcher
$19/ person / month
Go deeper from evidence to experiment design.
- All Explorer capabilities
- Compare findings and develop testable hypotheses
- Plan baselines, controls, metrics, and failure criteria
- Export research briefs and retain plan revisions
Lab
$79/ team / month
Keep a team’s evidence and decisions connected.
- Researcher capabilities for up to 5 seats
- Shared project workspaces and evidence collections
- Collaborative review of hypotheses and plans
- Team decision history and reusable research briefs
Proposed pricing. Features and usage allowances may change; contact us to discuss your needs.
A larger team or institution?
We’d like to understand your collaboration, integration, and research workflow needs.
06 / ABOUT EVIDENCETRELLIS
Room for ideas to grow.
A structure to build on.
EvidenceTrellis is an AI-native startup focused on research workflows. Our direction is an AI research copilot for evidence-backed planning.
The name reflects the idea: evidence provides the roots, and a trellis gives ideas a structure to grow. Researchers’ questions, evidence, and decisions are at the center of our work.
LET’S BUILD THE RIGHT NEXT STEP
Where does your
research get stuck?
We’d like to hear where your research process loses context, evidence, or time. If you work on research tools or have a workflow worth improving, let’s talk.
RESEARCH & PRODUCT CONVERSATIONSyifeishen@
evidencetrellis.dev
evidencetrellis.dev
Tell us about your research question, your current workflow, and the step you wish were easier.