Getting Started
Start with the question. Stay connected to the evidence.
Talvrin is designed to help you move from a market question to source-linked evidence, from evidence to context, from context to a research view, and from that view to ongoing monitoring. This guide explains the workflow and the boundaries that keep evidence inspectable.
Research and intelligence platform. No trade execution. No manufactured investment recommendations.
What Talvrin Is — And Is Not
Set the category before the task.
What It Is
- Source-linked research and monitoring for global public markets.
- A workspace connecting questions, evidence, context, research views, and monitoring.
- A system where source evidence remains independently inspectable.
- AI-assisted where appropriate, with provenance separation.
- Designed for self-directed investors, professionals, research teams, and institutions.
The Seven-Step Research Workflow
A research view should not end when an answer appears.
Talvrin’s model is continuous: Ask → Discover → Inspect → Understand → Build → Monitor → Reassess.
Ask
Frame a market, issuer, security, economic event, or policy question.
Start with a neutral, investigable question. Capture assumptions separately from evidence.
Discover
Find relevant evidence and context.
Use governed source discovery, prioritizing authoritative or appropriately licensed inputs where relevant.
Inspect
Open and review underlying sources.
Check source identity, original material, publication time, effective period, jurisdiction, version, and rights.
Understand
See how evidence relates to the question.
Distinguish support, contradiction, update, and context; keep source facts separate from interpretation.
Build
Develop and preserve a research view.
Connect conclusions and assumptions back to evidence so the view can be reviewed later.
Monitor
Continue watching evidence and assumptions.
Focus on meaningful changes, not notification volume.
Reassess
Return when the evidence base changes.
Review what changed, inspect the new source, and decide whether the research view requires revision.
Your First Research Task
A guided checklist for your first pass.
This checklist stands on its own — it does not depend on any specific product screen being available yet.
- 1
Define the question
Write one sentence describing what you are trying to understand, with market, entity, or event context where relevant.
Done when: you can restate the research question.
- 2
Identify source classes
List the kinds of evidence likely to matter: filings, economic releases, policy decisions, central-bank communications, market data, transcripts, or other governed sources.
Done when: you know not all sources carry equal authority.
- 3
Inspect timing and scope
Check publication time vs. effective period, jurisdiction, version, and supersession.
Done when: you can identify which date answers which question.
- 4
Separate evidence from interpretation
Mark what comes from the source, what Talvrin normalizes, what AI assists, and what you conclude.
Done when: you can explain the origin of each layer.
- 5
Build a view
Summarize the current evidence and assumptions without claiming certainty.
Done when: the view is traceable and reviewable.
- 6
Decide what to monitor
Identify the evidence or assumptions whose change would matter.
Done when: you have a bounded monitoring intent.
- 7
Reassess
When material evidence changes, reopen the source and update the view intentionally.
Done when: you understand research is ongoing, not one-shot.
Retention is earned by helping you build a durable research habit — not by streaks, pressure, or notification volume.
Understand the Evidence Chain
The source should never disappear behind the answer.
Where applicable, inspect who published the information, when, what period it applies to, which jurisdiction matters, whether a newer version supersedes it, and what access rights apply.
Source identity
Who published the information — a named organization, publisher, or issuer whenever known.
Source class
What kind of source it is, using governed classification only.
Publication time
When it was published, with timezone shown when material.
Effective / reference period
When the information applies — kept separate from publication time when they differ.
Version / supersession
Whether the document has changed, shown when revision history is supported.
Rights / access state
What you're permitted to access — never exposed beyond permitted use.
AI Principles for a First Session
AI can help you navigate the evidence. It does not become the evidence.
Generated interpretation should remain distinguishable from the underlying material you can inspect for yourself.
AI May Assist With
- Search and evidence discovery
- Source summarization
- Comparison
- Change identification
- Research organization
- Explanation and relationship mapping
- Contradiction surfacing
Choose a Starting Path
Optional routing, not a gate.
A path is a convenience — you can switch, skip, or ignore it. None of these imply availability beyond what’s confirmed.
Self-Directed Investor
Learn a disciplined way to investigate a market question.
Emphasis: Evidence vs. commentary, assumptions, uncertainty, source inspection, monitoring changes.
Explore Investment Professionals →Investment Professional
Move quickly from question to reviewable evidence.
Emphasis: Source authority, timing, jurisdiction, version, evidence relationship, reproducibility.
Explore the Platform →Research Team
Create research that remains reviewable and reusable.
Emphasis: Provenance, institutional memory, collaboration concepts, change monitoring.
Explore Research Teams →
Coverage and Availability Truth
Global architecture is not the same as universal live coverage.
Use current coverage information to understand what is supported, limited, planned, or not yet released.
Common First-Session Questions
Answers before you start.
What should I do first?
Start with a specific research question, identify the evidence that would support or challenge it, inspect the underlying sources, then preserve a view that can be reassessed when evidence changes.
Is Talvrin a trading platform?
No. Talvrin is designed for research and market intelligence, not trade execution.
Does Talvrin tell me what to buy or sell?
No. Talvrin is not a stock-tipping or manufactured recommendation engine; users remain responsible for their decisions.
Can I trust an AI summary as the source?
No. AI may assist research, while the underlying evidence remains separately inspectable and governed.
Why do publication date and effective period both matter?
A document may be published on one date but apply to another period; keeping them separate prevents false context.
Why does jurisdiction matter?
Regulation, policy, market structure, source rights, and economic meaning can differ across jurisdictions.
Does "global" mean every market is available?
No. Talvrin is architected globally, while released coverage is stated separately and progressively.
Have a different question? See the full FAQs.
Continue Learning
Where each Resources page picks up.
Getting Started teaches the workflow once. Everything else routes deeper without repeating it.
Help Center
NOT YET AVAILABLETask support, troubleshooting, recovery, and cross-resource discovery.
Getting Started
CURRENT PAGEOrientation and the first-workflow mental model.
Glossary
NOT YET AVAILABLEStable definitions for Talvrin and market/evidence terminology.
Release Notes
NOT YET AVAILABLEMaterial product changes and update history.
Build research that stays connected to the evidence.
And know where to look when that evidence changes.
Research and intelligence platform. No trade execution. No manufactured investment recommendations.