farnessPrototype

Policy futures · prototype

Forecasts on every consequential cell of government data

Farness analyst agents forecast published government statistics, law-encoded policy parameters, and outcomes conditional on policy states. Each cell carries a calibrated interval and an audit trail of the reasoning behind it.

Prototype status

Live product
Catalog, filters, routes, forecast pages, resolution rules.
Live API paths
CPI-U and two CTC cells stream through api.farness.ai.
Static mock traces
Most analyst reasoning and seeded estimates are prewritten prototype content.

All three forecast types.

Government dataStatic mock trace
resolves Sep 2028

SPM child poverty rate, 2027

forecast · 80% CI11.8%
10.5%11.8%13.4%
Government dataStatic mock trace
resolves Jan 2027

Unemployment rate, December 2026

forecast · 80% CI4.3%
3.9%4.3%4.8%
Government dataLive API path
resolves Jan 2027

CPI-U annual average inflation, 2026

forecast · 80% CI2.6%
2.1%2.6%3.2%
Government dataStatic mock trace
resolves Sep 2027

Median household income, 2026

forecast · 80% CI$81,200
$79,100$81,200$83,400
Government dataStatic mock trace
resolves Oct 2027

Federal individual income tax revenue, FY2027

forecast · 80% CI$2,740
$2,580$2,740$2,910
Government dataStatic mock trace
resolves Jan 2027

Real GDP growth, 2026 (Q4/Q4)

forecast · 80% CI+1.9%
+0.9%+1.9%+2.8%
Government dataStatic mock trace
resolves Sep 2027

Uninsured rate (under 65), 2026

forecast · 80% CI10.1%
9.2%10.1%11.2%
Government dataStatic mock trace
resolves Jan 2027

Labor force participation rate, December 2026

forecast · 80% CI62.4%
61.9%62.4%62.8%
PolicyStatic mock trace
resolves Apr 2027

CTC monthly maximum per child, TY2027

forecast · 80% CI$183/mo
$83/mo$183/mo$250/mo
PolicyLive API path
resolves Dec 2027

Cost of a $3,000 fully refundable CTC, TY2026

forecast · 80% CI$109.6B
$78.8B$109.6B$140.4B
PolicyLive API path
resolves Aug 2028

CTC outlays under current law, TY2026

forecast · 80% CI$60.5B
$52B$60.5B$70B
PolicyStatic mock trace
resolves Aug 2028

EITC outlays, TY2026

forecast · 80% CI$74B
$65B$74B$84B
PolicyStatic mock trace
resolves Dec 2028

Revenue cost of a $40k SALT cap, TY2027

forecast · 80% CI$38B
$25B$38B$55B
PolicyStatic mock trace
resolves Jan 2027

Federal minimum wage on Jan 1, 2027

forecast · 80% CI$7.25
$7.25$7.25$9.50
PolicyStatic mock trace
resolves Dec 2027

Standard deduction, joint filers, TY2027

forecast · 80% CI$32,000
$16,000$32,000$35,000
PolicyStatic mock trace
resolves Dec 2027

Top individual income tax rate, TY2027

forecast · 80% CI37.0%
37.0%37.0%39.6%
PolicyStatic mock trace
resolves Dec 2027

SALT deduction cap, TY2027

forecast · 80% CI$22,000
$10,000$22,000$80,000
PolicyStatic mock trace
resolves Oct 2026

SNAP maximum allotment, family of 4, FY2027

forecast · 80% CI$1,010
$975.00$1,010$1,045
Government dataStatic mock trace
resolves Nov 2027

SNAP benefit outlays, FY2027

forecast · 80% CI$128B
$116B$128B$143B
Government dataStatic mock trace
resolves Nov 2027

ACA premium tax credit outlays, FY2027

forecast · 80% CI$94B
$62B$94B$132B
Government dataStatic mock trace
resolves Jun 2028

Medicaid and CHIP enrollment, Dec 2027

forecast · 80% CI82M
77M82M88M
Government dataStatic mock trace
resolves Sep 2027

SPM poverty rate, 2026

forecast · 80% CI12.4%
11.2%12.4%13.9%
Government dataStatic mock trace
resolves Nov 2027

Unemployment insurance outlays, FY2027

forecast · 80% CI$49B
$32B$49B$78B
Government dataStatic mock trace
resolves Nov 2027

SSI federal payments, FY2027

forecast · 80% CI$75B
$68B$75B$84B
Government dataStatic mock trace
resolves Nov 2027

Federal Medicaid outlays, FY2027

forecast · 80% CI$655B
$600B$655B$720B
Government dataStatic mock trace
resolves Aug 2028

Children receiving the CTC, TY2026

forecast · 80% CI48M
45M48M51M
Government dataStatic mock trace
resolves Nov 2027

Housing Choice Voucher outlays, FY2027

forecast · 80% CI$36B
$32B$36B$42B
Government dataStatic mock trace
resolves Oct 2027

Federal payroll tax receipts, FY2027

forecast · 80% CI$1,860B
$1,760B$1,860B$1,985B
Government dataStatic mock trace
resolves Nov 2027

Social Security OASDI benefit outlays, FY2027

forecast · 80% CI$1,605B
$1,540B$1,605B$1,685B
PolicyStatic mock trace
resolves Aug 2028

Refundable AOTC outlays, TY2026

forecast · 80% CI$6.2B
$4.8B$6.2B$7.8B
Government dataStatic mock trace
resolves Mar 2028

WIC average monthly participation, FY2027

forecast · 80% CI6.9M
6.3M6.9M7.6M
Government dataStatic mock trace
resolves Mar 2028

National School Lunch participation, SY2026-27

forecast · 80% CI30.4M
28.8M30.4M32.3M
Government dataStatic mock trace
resolves Apr 2027

ACA exchange plan selections, OEP 2027

forecast · 80% CI19.8M
14M19.8M25M
Government dataStatic mock trace
resolves Nov 2027

Medicare benefit outlays, FY2027

forecast · 80% CI$1,110B
$1,035B$1,110B$1,210B
Government dataStatic mock trace
resolves Mar 2028

TANF federal outlays, FY2027

forecast · 80% CI$17B
$15B$17B$19.5B
Government dataStatic mock trace
resolves Mar 2028

Child Care and Development Fund outlays, FY2027

forecast · 80% CI$12.5B
$9B$12.5B$17B
Government dataStatic mock trace
resolves Oct 2027

Estate and gift tax receipts, FY2027

forecast · 80% CI$45B
$28B$45B$72B
Government dataStatic mock trace
resolves Dec 2029

EITC claimant returns, TY2027

forecast · 80% CI23.4M
21.6M23.4M25.5M
Government dataStatic mock trace
resolves Oct 2027

Corporate income tax receipts, FY2027

forecast · 80% CI$570B
$430B$570B$730B
Government dataStatic mock trace
resolves Mar 2028

SNAP average monthly participation, FY2027

forecast · 80% CI40.6M
37.4M40.6M44.1M
Government dataStatic mock trace
resolves Feb 2029

Pell Grant recipients, award year 2027-28

forecast · 80% CI7.2M
6.4M7.2M8.2M
Government dataStatic mock trace
resolves Jun 2028

Head Start funded enrollment, FY2027

forecast · 80% CI0.8M
0.7M0.8M0.9M
Government dataStatic mock trace
resolves Jun 2028

TANF average monthly families, FY2027

forecast · 80% CI0.7M
0.6M0.7M0.9M
Government dataStatic mock trace
resolves Mar 2029

CCDF average monthly children served, FY2027

forecast · 80% CI1.4M
1.1M1.4M1.8M
Government dataStatic mock trace
resolves Sep 2029

LIHEAP households assisted, FY2027

forecast · 80% CI5.1M
3.4M5.1M6.9M
Government dataStatic mock trace
resolves Mar 2028

School Breakfast participation, SY2026-27

forecast · 80% CI15.2M
14M15.2M16.7M
Government dataStatic mock trace
resolves Feb 2028

SSI recipients, December 2027

forecast · 80% CI7.4M
7.1M7.4M7.7M
Government dataStatic mock trace
resolves Mar 2028

SNAP average monthly households, FY2027

forecast · 80% CI21.7M
19.8M21.7M23.8M
Government dataStatic mock trace
resolves Apr 2028

Medicaid and CHIP child enrollment, December 2027

forecast · 80% CI36.8M
34.5M36.8M39.4M
Government dataStatic mock trace
resolves Jun 2028

Housing Choice Voucher households leased, December 2027

forecast · 80% CI2.3M
2.2M2.3M2.5M
Government dataStatic mock trace
resolves Mar 2028

School Breakfast federal outlays, FY2027

forecast · 80% CI$6.7B
$5.8B$6.7B$7.8B
Government dataStatic mock trace
resolves Dec 2028

Child support distributed collections, FY2027

forecast · 80% CI$27.8B
$25B$27.8B$31B
Government dataStatic mock trace
resolves Jun 2028

Federal Pell Grant outlays, FY2027

forecast · 80% CI$34.5B
$28B$34.5B$43B
Government dataStatic mock trace
resolves Mar 2028

WIC federal outlays, FY2027

forecast · 80% CI$8.5B
$7.2B$8.5B$10B
Government dataStatic mock trace
resolves Mar 2028

LIHEAP federal funding, FY2027

forecast · 80% CI$4.1B
$2.4B$4.1B$6.8B
Government dataStatic mock trace
resolves Feb 2028

OASDI beneficiaries, December 2027

forecast · 80% CI70M
68.8M70M71.4M
Government dataStatic mock trace
resolves Dec 2028

Child and dependent care credit returns, TY2026

forecast · 80% CI5.9M
5M5.9M7M
Government dataStatic mock trace
resolves Dec 2028

Premium tax credit claimant returns, TY2026

forecast · 80% CI10.8M
7.4M10.8M15.2M
Government dataStatic mock trace
resolves Dec 2028

Additional Child Tax Credit returns, TY2026

forecast · 80% CI16.2M
13.2M16.2M19.6M
Government dataStatic mock trace
resolves Dec 2028

Itemized charitable contributions, TY2026

forecast · 80% CI$305B
$245B$305B$375B
Government dataStatic mock trace
resolves Mar 2028

Medicare total enrollment, December 2027

forecast · 80% CI70.7M
69.2M70.7M72.3M
Government dataStatic mock trace
resolves Feb 2028

Retired-worker average benefit, December 2027

forecast · 80% CI$2,190/mo
$2,110/mo$2,190/mo$2,285/mo
PolicyStatic mock trace
resolves Dec 2027

Refundable CTC amount, TY2027

forecast · 80% CI$1,800
$1,000$1,800$2,500
PolicyStatic mock trace
resolves Oct 2026

Maximum EITC, 3+ children, TY2027

forecast · 80% CI$8,500
$8,050$8,500$9,100
Government dataStatic mock trace
resolves Mar 2028

SNAP average benefit per person, FY2027

forecast · 80% CI$232/mo
$216/mo$232/mo$252/mo
Government dataStatic mock trace
resolves Jun 2028

ACA PTC average subsidy per enrollee, FY2027

forecast · 80% CI$690/mo
$560/mo$690/mo$860/mo
ConditionalStatic mock trace
resolves Sep 2029

Child poverty 2028 | TCJA extension passes by Q2 2026

forecast · 80% CI11.5%
10.3%11.5%12.8%
ConditionalStatic mock trace
resolves Sep 2027

Child poverty 2026 | $3,000 fully refundable CTC

forecast · 80% CI7.2%
5.8%7.2%9.0%
ConditionalStatic mock trace
resolves Oct 2028

Federal individual income tax revenue FY2028 | SALT cap fully repealed

forecast · 80% CI$2,790
$2,630$2,790$2,960
ConditionalStatic mock trace
resolves Sep 2029

Uninsured rate (under 65) 2028 | ACA enhanced subsidies expire

forecast · 80% CI11.3%
10.2%11.3%12.4%
Decision framework for agents

Arm your agent’s decisions with forecasting.

farness intercepts agent decisions and demands a forecast: a KPI, a confidence interval, a base rate, disconfirming evidence, and a review date. Works with Codex, Claude Code, and any agent that speaks MCP.

Get startedRead the paper
Decision prompt:
Should we rewrite the auth layer now?
Reframed as:
P(critical auth incidents decrease by >40% in 90 days | rewrite now)
KPI:
critical_auth_incidents / 90d
Forecast:
rewrite now:
58% [42-71]
defer 60d:
31% [19-44]
Base rate:
27%similar infra rewrites yielding material reliability gains
Disconfirming evidence:
ops fixes may solve this fasterrewrite could slip roadmap deliveryrecent outage may overweight urgency
Review date:
2026-06-15
How farness works

From intuition to instrument

01

Intercept

Catch decision-language before the model hardens into advice. When a prompt sounds like 'Should we...?' or 'Which is better?', farness reframes it as a forecastable choice.

02

Reframe

Convert vague 'Should I?' into explicit, measurable outcome questions. Define the KPIs that would actually tell you whether the decision was good.

03

Anchor

Produce numeric forecasts with confidence intervals, reference classes from comparable situations, disconfirming evidence, and a review date for accountability.

Workflow demo

Watch the packaged path end to end

The clip below shows the current Codex path exactly the way the docs describe it: install the package, register the local MCP server, use $farness in Codex, then pull the decision back out of the local store.

1
python -m pip install 'farness[mcp]'
2
farness setup codex
3
$farness inside Codex, then review the saved decision locally
Open install docsView source

Rendered from a real Codex session using the local farness skill and MCP server, then exported as a clean 4K terminal demo.

Open 4K MP4Open poster
From intuition to forecast

What the framework forces into view

Diffuse prompt

“Should I refactor this module first?”

Farness output
KPI: bug_rate / 30d
Event:>25% bug reduction
Horizon: 90 days
Forecast: 44% [28-61]
Base rate: 22%
Disconfirming evidence: migration drag, auth edge cases
Review: 2026-06-15
Research

Stability-under-probing

11
study 1 scenarios
2
studies in the paper
8
held-out validation cases

The paper introduces stability-under-probing as a way to evaluate decision prompts without waiting for outcomes. In Study 1, farness looked more prepared for the shared probe battery on Claude Opus 4.6 and GPT-5.4.

Study 2 then added held-out probes and showed the broader claim weakens sharply off-framework. That makes the paper a methods result first, not proof that farness is universally superior.

The useful claim is narrower and better: structured decision prompts can be tested empirically, and farness is one case study.

Read the full paper
Output primitives

What farness produces

KPI

What outcome actually matters. Defined before the analysis, not after.

Forecast

Numeric probability for each option. Not opinions — predictions you can score.

Confidence interval

The honest range around the estimate. Calibrated uncertainty, not false precision.

Base rate

What usually happens in comparable situations. The outside view as empirical anchor.

Disconfirming evidence

What counter-evidence, failure modes, or decision traps could make the leading option wrong.

Review date

When to check the forecast against reality. Accountability built in.

AI is often fluent about decisions before it is rigorous about them. farness adds structure before confidence hardens into action.

Agent integrations

Use it natively or from the CLI

Farness now has a package-first agent path: a local MCP server for persistence, packaged skills for Codex and Claude Code, and the same forecast structure used in the paper. The Claude plugin remains optional, and the CLI is a local store and calibration surface, not an LLM client. If setup drifts, `farness doctor --fix` repairs the local integration.

Codex

Install the package, run one setup command, then use $farness when a decision prompt shows up.

$ python -m pip install 'farness[mcp]'
$ farness setup codex
$ # restart Codex, then use $farness
Claude Code

Use the same single-command setup flow for Claude. The plugin is still available if you prefer slash-command UX.

$ python -m pip install 'farness[mcp]'
$ farness setup claude
$ # restart Claude Code
CLI / Python

Local decision log and calibration tool. No LLM API key required unless you run separate experiment code against external models.

$ python -m pip install farness
$ farness new "Should we rewrite the auth layer?"
$ farness calibration
View on GitHubopen docsAgent workflows

See further before you decide.

Start with farness