Inspect the evidence behind investor belief data.
BeliefState's REST API and MCP tools return structured beliefs, exact evidence, revisions, risks, uncertainty, and measured outcomes. These public records show the data with its proof attached.
Updated: today
Investor belief data should come with receipts.
Markets remember winners and forget old screenshots, deleted posts, changed targets, and losing periods. BeliefState keeps the source trail intact so reputation can be checked against evidence.
Public pages show only records approved for public release. Private research, revisions, and workspace decisions remain private.
- Public trade ideas
- 52
- Tickers covered
- 29
- Investors attributed
- 9
- Resolved windows
- 126
Four questions
What every useful trade idea should answer.
01
Who said it first?
Author, source URL, and publication time stay attached to the call. Attribution does not disappear when an idea spreads.
02
What did they actually say?
Each record shows the exact evidence behind the extracted ticker, action, direction, target, horizon, and invalidation condition.
03
What would prove it wrong?
Risk and counterevidence belong beside the thesis. BeliefState keeps revisions instead of replacing an old claim with a cleaner story.
04
What happened next?
Outcomes compare the security with a benchmark across fixed market-session windows. Losing and pending periods remain visible.
Start with what you know
Find the idea by claim, ticker, or investor.
Trade ideas
Read the original claim
Inspect the exact quote, source, timestamp, direction, target, invalidation condition, and every outcome window.
52 public records
Open record →Tickers
See the research around a company
Compare authors, direction, timing, stated risk, and market-relative results for one security.
29 tickers
Open record →Investors
Check the record behind the reputation
See which calls are attributed to an investor, where the evidence came from, and how resolved windows performed.
9 public investors
Open record →Current public evidence
Latest source-linked trade ideas.
shorting $AAOI on the pop, one year of bottleneck on a commodity product does not make a great company
- Call
- short · bearish
- Time horizon
- Not stated
- Target
- Not stated
- Primary result
- 5 sessions · — pending
Exact source evidence
“shorting $AAOI on the pop, one year of bottleneck on a commodity product does not make a great company”
Market-relative outcomes
Direction-adjusted return versus benchmark. Losses stay in the record.
1 sessions
+5.2%
working
5 sessions
—
pending
21 sessions
—
pending
42 sessions
—
pending
63 sessions
—
pending
126 sessions
—
pending
252 sessions
—
pending
504 sessions
—
pending
Method heuristic-explicit-call-v1 · Evidence extraction confidence 65%
SOLD $FLNC @13.97
- Call
- sell · bearish
- Time horizon
- Not stated
- Target
- Not stated
- Primary result
- 5 sessions · — pending
Exact source evidence
“SOLD $FLNC @13.97”
Market-relative outcomes
Direction-adjusted return versus benchmark. Losses stay in the record.
1 sessions
—
pending
5 sessions
—
pending
21 sessions
—
pending
42 sessions
—
pending
63 sessions
—
pending
126 sessions
—
pending
252 sessions
—
pending
504 sessions
—
pending
Method kevin-xu-ytd-user-reviewed-v1 · Evidence extraction confidence 100%
Currently all in $FLNC in my challenge account
- Call
- buy · bullish
- Time horizon
- Not stated
- Target
- Not stated
- Primary result
- 5 sessions · — pending
Exact source evidence
“Currently all in $FLNC in my challenge account”
Market-relative outcomes
Direction-adjusted return versus benchmark. Losses stay in the record.
1 sessions
—
pending
5 sessions
—
pending
21 sessions
—
pending
42 sessions
—
pending
63 sessions
—
pending
126 sessions
—
pending
252 sessions
—
pending
504 sessions
—
pending
Method kevin-xu-ytd-user-reviewed-v1 · Evidence extraction confidence 100%
BUY AMD: Primary post frames AMD AH selloff after clean beat as a potential GOOGL moment that soared ~20% next few days, while quoting author's own post that labeled double-beat selloffs a no-brainer buying opportunity; reinforced by same-day posts calling AMD an easy buy that could rocket like PLTR.
- Call
- buy · bullish
- Time horizon
- the next few days
- Target
- Not stated
- Primary result
- 5 sessions · — pending
Exact source evidence
“$AMD down after hours despite a clean beat / Q3 guidance above expectations.”
Market-relative outcomes
Direction-adjusted return versus benchmark. Losses stay in the record.
1 sessions
+0.3%
flat
5 sessions
—
pending
21 sessions
—
pending
42 sessions
—
pending
63 sessions
—
pending
126 sessions
—
pending
252 sessions
—
pending
504 sessions
—
pending
Method ai-implied-call-v1 · Evidence extraction confidence 74%
BUY SPCX: Primary post explicitly tells shorts to cover SPCX while quoting Musk’s warning; nearby same-author posts supply high SI, institutional gross-up, and sandbagged-earnings context that jointly imply a long/cover-short stance.
- Call
- buy · bullish
- Time horizon
- near-term into/around lockup and short-cover dynamics
- Target
- Not stated
- Primary result
- 5 sessions · — pending
Exact source evidence
“$SPCX: shorts better cover 😂”
Market-relative outcomes
Direction-adjusted return versus benchmark. Losses stay in the record.
1 sessions
+1.0%
flat
5 sessions
—
pending
21 sessions
—
pending
42 sessions
—
pending
63 sessions
—
pending
126 sessions
—
pending
252 sessions
—
pending
504 sessions
—
pending
Method ai-implied-call-v1 · Evidence extraction confidence 75%
BUY CCXI: Author explicitly reports starting a position after Wave-3 identification, breakout clearance and multiple new buy-signals, while noting potential to get aggressive and not wanting to miss the move.
- Call
- buy · bullish
- Time horizon
- Not stated
- Target
- Not stated
- Primary result
- 5 sessions · — pending
Exact source evidence
“the start of Wave 3 for $CCXI”
Market-relative outcomes
Direction-adjusted return versus benchmark. Losses stay in the record.
1 sessions
-1.4%
flat
5 sessions
—
pending
21 sessions
—
pending
42 sessions
—
pending
63 sessions
—
pending
126 sessions
—
pending
252 sessions
—
pending
504 sessions
—
pending
Method ai-implied-call-v1 · Evidence extraction confidence 85%