agent gate
read path
The agent can inspect policy-safe markets and existing thesis records before it writes anything.
@evapredicts campaign · agent forecast interface
forecast agents need an interface for receipts.
If an agent can write a market call, the product question is not whether it sounds smart. It is whether the user can inspect the sources, revision trigger, author trail, and publish boundary after the feed moves on.
Campaign hypothesis
agent builders will click when the ask is interface quality, not prediction magic.
The audience already believes agents can generate forecasts. The sharper wedge is that public forecast agents need a constrained interface: read before writing, prepare before publishing, receipts before distribution.
interface checks
four gates before an agent forecast gets amplified.
agent gate
draft boundary
The agent prepares a thesis or revision for review. It does not silently publish, trade, or broadcast.
agent gate
receipt fields
The output carries source fit, revision trigger, author identity, wallet source, and a clear publish-state ceiling.
agent gate
public object
The user gets a thesis page or proof link that survives the feed instead of trusting a one-shot answer.
target audience
builders who need agent outputs to be inspectable.
agent builders turning prediction-market data into public forecast outputs
Show the interface boundary, then route them to a proof-backed thesis where the standard is visible.
prediction-market operators who want agent distribution without unverifiable screenshots
Show the interface boundary, then route them to a proof-backed thesis where the standard is visible.
crypto analysts using agents to draft, counter, or revise market theses
Show the interface boundary, then route them to a proof-backed thesis where the standard is visible.
protocol teams comparing AI forecast UX by source quality and permission boundaries
Show the interface boundary, then route them to a proof-backed thesis where the standard is visible.
Campaign sequence
make agent-readiness measurable before widening @evapredicts distribution.
- Route agent-builder traffic to this page with utm_campaign=agent_forecast_interface.
- Send high-intent readers to the agent manifest, MCP quickstart, proof thesis, or compose flow depending on their intent.
- Keep the wedge only if it produces manifest opens, quickstart clicks, proof-record reads, compose starts, author-record clicks, or follows.
@evapredicts copy to approve
ship the boundary, not an agent hype post.
External posting still needs explicit approval. Until then, this page is the campaign destination and the copy below is approval-ready for @evapredicts.
forecast agents do not need more confidence.
they need an interface for receipts.
what market did the agent inspect? what source would change the thesis? who authored it? what did it actually publish, if anything?
Eva's wedge is the proof object around the forecast, not a louder answer machine.
start here: https://eva.jaack.me/campaigns/agent-forecast-interface?utm_source=x&utm_medium=social&utm_campaign=agent_forecast_interface&utm_content=interface_post
Metric to watch: sessions with utm_campaign=agent_forecast_interface, agent-manifest opens, MCP quickstart clicks, proof-record reads, compose starts, author-record clicks, and @evapredicts follow clicks. Do not claim traction until those are measured.