Resources

Getting Started with Taken: Markets, Agents and Automations

A practical orientation to the current Taken workspace: explore an instrument, ask a research question, build an agent and test a visual workflow.

Taken organizes financial research around Markets, Agents and Automations. You do not need to configure everything at once. Start with one instrument and a question you can verify, then turn the parts you repeat into a more structured process.

This guide describes the current workspace rather than promising universal data coverage or automatic investment decisions. Check the actual status of an instrument or connected tool whenever you use it.

Begin in Markets

Use the central search to identify a stock, ETF, supported futures contract, crypto asset or meme coin. Check the result's identity, particularly when symbols overlap. For decentralized tokens, the network and contract details matter; for futures, the selected contract matters.

Open the instrument and review the available chart, observation period and quote timestamp. A previous-session quote is not an intraday observation. Missing data should remain visibly unavailable rather than being interpreted as a zero value.

The Markets board brings market panels and community research into one workspace. Arrange the available panels around the task you are doing, and resize them as needed. The goal is a useful view of the evidence, not the largest possible number of widgets.

Ask a focused research question

Use Taken's contextual AI for an instrument-specific question or a broader research task. State the time horizon and ask for sources where current information is required. Read the underlying material before relying on a numerical or causal claim.

Try asking for a structured answer: observations, interpretation and unresolved questions. That makes it easier to tell whether the response has actually answered your question or merely produced a plausible summary.

Build an agent for a task you repeat

Open Agents when you have a repeatable objective, such as reviewing available company results with the same checklist. Describe the task, inspect the generated configuration and test the saved version.

Review sources, instructions, constraints and actions. An agent should have only the resources needed for its purpose. A missing connection or unsupported dataset needs a real resolution, not an instruction to guess. The first-agent walkthrough explains the review process in more detail.

Use Automations for a sequence

Open Automations to connect steps on the visual canvas. A workflow can organize a trigger, data retrieval, agent work, conditions and outputs where those actions are supported.

Keep your first workflow manual and small. Inspect each node, test the saved configuration and read actual step results. Add a schedule or connected delivery destination after the basic path works. A diagram that looks connected is not proof of a completed run.

Keep private and public work distinct

Community research can help you discover a new question or reusable process. Review public claims independently. When sharing your own work, remove personal financial information and make the research context clear.

A public agent or automation should expose reusable configuration rather than account credentials. Another user needs their own connections and permissions before running a copy. Explore social investing for a practical approach to reading and sharing research.

Finish your first session with one saved question, one checked source and one next step. Expanding a process you understand is easier than debugging a large collection of unfamiliar tools.

Educational information, not a recommendation to buy or sell an investment. Data availability and product capabilities can change. Examples in this article are illustrative unless identified as observed data.

Put your research to work

Explore it in Taken.

Markets, research agents and visual workflows in one workspace.

Get started with Taken