Public Intelligence Overview
Public Intelligence is the TradeOS workflow for turning crypto evidence into human-readable research and source-grounded intelligence products. It is designed for thesis review, alerting, feedback, and dry-run publishing before any content is sent to public platforms.
The goal is to help readers understand what TradeOS is seeing without requiring them to be quant researchers or protocol insiders. A good public thesis should make the background, token identity, supporting evidence, risk, uncertainty, and material-change triggers understandable to traders, long-term investors, builders, and curious people.
Standalone public intelligence is not a managed trading signal service and it cannot guarantee that an asset is safe. A public thesis explains what TradeOS is watching, why it may matter, what the supporting evidence says, what is uncertain, what could be risky, and what would change the view. The flagship builder pattern is different: individual private-use, self-hosted cockpits can turn that evidence into buy, sell, trim, avoid, watch, or pass recommendations when they run in the operator's own environment with the operator's rules, keys, approvals, and execution controls.
The reason this matters is that agents and builder products are only useful when their market context is fresh, source-backed, identity-aware, and reviewable. See Problem Space for the cost of missing TradeOS-like intelligence: wrong-token routing, stale answers, thin evidence, unsafe caveat-free recommendations, missed watchlist changes, and weak audit trails. The same page includes magnitude anchors so readers can translate the gap into exposed capital, manual review cost, and builder trust risk without treating TradeOS as a guaranteed loss-prevention product.
Public Value Ladder
TradeOS keeps the public story simple:
Use TradeOS free.
Earn Data Intel Credits by improving intelligence quality.
Build and earn on public intelligence.
Pay when you need private intelligence, scale, alerts, automation, or data rights.
Public Intelligence is the evidence layer behind that ladder. Free readers get bounded public context. Useful feedback can improve the review loop and earn scoped Data Intel Credits for temporary dashboard depth, public Ask packs, and read-only Review Lab access. Builders can package public evidence into paid services, workflows, agents, dashboards, vertical apps, and research products while preserving target IDs and provenance; builder feedback affects app reputation and quota confidence rather than becoming personal GUI credit. Paid TradeOS starts when a workflow needs private intelligence products, production volume, delivery, automation, machine access, or explicit data rights.
The builder flywheel is the compounding effect behind the public kit: TradeOS evidence helps builders ship useful paid products; those products create usage, feedback, and outcome labels; feedback improves TradeOS coverage, ranking, caveats, and freshness; better intelligence makes builder products stronger; successful workflows then move to paid TradeOS, x402, or enterprise access.
Public API usage is bounded the same way. Anonymous reads stay small; verified builder keys receive a 7-day starter window; useful attributed feedback can refresh public depth; serious projects can request review; machine-scale reads, alerts, exports, replay, private intelligence context, and data rights require x402 payment, paid API access, or entitlement.
Active defaults remain bounded but now favor attributed builders: anonymous API preview is 2/minute, 40/day, and 5 symbols/day; baseline app keys are 5/minute, 100/day, and 10 symbols/day; starter or earned app keys are 15/minute, 400/day, and 30 symbols/day. GUI Ask TradeOS still starts at 3 anonymous questions or 10 signed-in starter questions.
What the workflow produces
The publisher currently supports these human-facing outputs:
| Output | Purpose |
|---|---|
| Thesis Watchlist Pulse | Daily ranked intelligence watchlist showing the strongest and weakest active research setups |
| New Thesis Candidates | Daily candidate theses that may deserve a public research draft |
| Material Change Alerts | Daily alerts when source evidence changes enough to affect an active thesis |
| Token Risk Digest | Weekly review of liquidity, sellability, contract, and identity risk |
| Outcome Follow-Up | Weekly review of resolved, stale, corrected, or invalidated theses |
| Thesis Checkpoints | Monthly checkpoint candidates for longer-running active theses |
| Narrative Radar | Monthly digest for sector rotation and long-horizon narrative drift |
How a draft is built
The service follows a source-first process:
- TradeOS fetches structured inputs from approved source adapters.
- The inputs are normalized into an evidence pack with source, timestamp, claim, freshness, confidence, chain, and contract context where available.
- Deterministic templates create the first draft and enforce required sections.
- The LLM may rewrite the draft into clearer prose, but it must stay inside the evidence pack.
- Policy gates check source coverage, claim safety, freshness, duplication, channel limits, and dry-run state.
- The result is sent for operator review by email and stored as a dry-run platform publication.
- Review decisions, corrections, material changes, and later outcomes are recorded for follow-up.
The practical rule is simple: no source, no claim.
Feedback and follow-up
Public Intelligence is not a one-shot writing workflow. TradeOS tracks whether a thesis is strengthened, weakened, invalidated, stale, unresolved, or materially changed. That memory helps the system improve future watchlists, follow-ups, evidence ranking, claim framing, and operator review queues.
This is how the public research layer stays honest: a thesis should not only sound good when published. It should remain connected to what happened later.
The app also exposes a Data Intel Credit loop. Human users can label specific
cards and review tasks to earn account-based DTI credits after quality checks.
Builders using the distribution kit can submit feedback with provenance fields
so TradeOS can distinguish direct human judgment, human-assisted review, agent
feedback, and automation. Human DTI is not API-convertible; attributed builder
feedback contributes to a visible app_reputation_dti summary for the app key
after Feedback Ops approval. That summary is non-personal, non-convertible, and
does not unlock paid capacity. Builders can inspect app-key feedback activity
from Builder/App Feedback, while Developer Keys remains the key-management
detail page. Agents can self-check their app-key feedback status with the
public-intel app key. See Data Intel Credit Loop.
Builders can start from the open-source TradeOS Public Intelligence Kit to build bots, dashboards, SDK integrations, MCP tools, agent workflows, and research products on top of TradeOS public evidence. See Earn as a Builder Kit for the repo, setup path, how builders earn, feedback-provenance rules, Venice AI pattern, Venice AI subscription, paid boundaries, and contribution guidance.
The clearest consumer-facing product shape is a private self-hosted Symbol Cockpit:
symbol -> TradeOS evidence -> local recommendation -> local decision -> feedback
Builders can package or extend that cockpit for users to self-host while the runtime calls TradeOS for public or paid intelligence. Local watchlists, wallet context, strategy notes, bot rules, exchange keys, approvals, execution adapters, and logs should remain local unless the operator chooses to send feedback or authenticated context to TradeOS. That is the flagship split: TradeOS is the intelligence product layer; the cockpit is the self-hosted operator's decision runtime.
The bridge from cockpit recommendation to action workflow is a non-executable action intent. It preserves evidence and risk context while requiring the self-hosted operator to choose venue, account, size, order type, timing, approval, custody, and executor locally. It is not an order or transaction.
Why chain and contract identity matter
Crypto symbols are not unique. The same ticker can appear on multiple chains, and fake or scam tokens can imitate a legitimate project. Public thesis outputs therefore include chain and contract identity where available, so readers can tell which asset TradeOS is referring to.
This is especially important for meme, Base, AI, DeFi, and long-tail assets where symbol collisions are common.
Review before publishing
The dry-run setup exists so public writing can be tuned before it reaches platforms:
- email notifications show the full review copy;
- platform output can be reviewed as a draft before live publishing is enabled;
- operators can pause publishing or use the kill switch if the workflow behaves unexpectedly.
Public intelligence should be readable, source-backed, and honest about uncertainty. Standalone public writing should not sound like an internal dump or pretend to know a reader's private sizing, portfolio, or execution constraints. Actionable buy/sell/trim/avoid language belongs in the private self-hosted Symbol Cockpit or other operator-owned workflows that make the evidence, assumptions, local controls, and execution boundary visible.