Verifiable research infrastructure · Live

Deep Research
as Infrastructure.

AI agents need trustworthy external knowledge. Deep Water transforms hundreds of sources into verified evidence. Reports are generated from evidence, not from raw web pages.

Evidence first · Text last · The frontier model never reads raw sources

admin.deepwater.live
Deep Water
U

Provider input units

4.0M

immutable provider metering

Jobs this month

47

14 Deep, 28 Standard, 5 Light

URLs fetched

1,962

avg 41.7 per job

Provider actual cost

$1.18

centrally recorded

Token spend · last 8 weeks

Weekly Tokens

Recent research

GLP-1 cardiovascular outcomes
8 tok
Kubernetes 1.33 breaking changes
3 tok
EU AI Act Annex III scope
15 tok
Competitor SOC2 Type II status
1 tok

The Question

Can you prove it's true?

Confident isn't the same as correct. An AI answer that sounds certain is not the same as one you can verify against its source.

The Problem

AI agents without research are coding blind.

Before an agent writes a line of code, it has to understand the world it's writing into. That knowledge is scattered across hundreds of sources — and a model alone cannot reach it reliably.

Documentation

Benchmarks

Incidents

Postmortems

RFCs

A larger model doesn't fix this. The bottleneck isn't reasoning — it's collecting and verifying the evidence the reasoning runs on.

Why Existing Solutions Fall Short

Generation is solved. Trust is not.

Report-first systems read the web, then write prose in one pass. The text reads well — but the evidence behind it is gone by the time you see it. Three problems follow from that.

01

You cannot verify outputs.

A report-first model blends what it read with what it already believed. There is no record of which source supports which claim — so you cannot tell a sourced fact from a fluent guess.

02

You cannot control the process.

A chat box hides every decision: which sources were read, how deep it searched, what it discarded. You get a finished answer and no way for software to inspect, gate, or reproduce how it was reached.

03

You cannot afford it at scale.

Pushing hundreds of raw pages through a frontier model on every run is expensive by construction. Run it across an agent fleet and the economics break before the quality does.

The report is the symptom, not the unit. When the report is generated directly from web content, there is nothing underneath it to audit. Deep Water inverts that order.

The Atomic Unit of Truth

Evidence Cards.

Not a citation feature. The building block the entire report is assembled from. Every card is one fact, bound to the exact text that supports it.

evidence_card #1182 confidence 96%

Fact

MCP authentication failures often stem from improper token-refresh handling rather than malformed credentials.

Source span

… clients MUST refresh the access token before expiry; failing to handle token refresh is the most common cause of authorization errors in long-running sessions …

Citation

Anthropic Documentation · lines 182–201

Verified against span

Auditability

Open any claim in the report and see the exact card, span, and source it was built from. Nothing is asserted without a receipt.

Traceability

Every sentence in the report walks back through a card to a span on a real page. No orphaned statements, no untraceable summaries.

Compliance

An evidence trail you can hand to a reviewer, an auditor, or a compliance team. Confidence scores and source provenance travel with the output.

Explainability

When a downstream agent acts on a fact, you can answer why — the card carries the reasoning chain from raw text to verified claim.

Evidence-First Architecture

Evidence first. Text last.

The report is generated from verified evidence — never directly from raw web pages. The frontier model is the last step in the pipeline, not the first.

Sources

Hundreds to thousands of URLs — docs, papers, filings, forums — searched and fetched.

Evidence Cards

Inexpensive models extract atomic facts, each bound to its exact source span.

Verification

Each card is checked against its span. Unsupported claims are dropped; conflicts are flagged.

Report

The frontier model writes from verified evidence only — one call, fully traceable.

The frontier model never reads raw sources.

It sees a clean, verified evidence set — not hundreds of noisy web pages. That's what makes the output auditable and the cost predictable.

Every claim traces back to a span.

Because text is assembled from evidence rather than generated over it, every sentence in the report walks back to the source it came from.

The Pipeline

Four phases. One API call.

A research pipeline built for auditability, traceability, and cost efficiency. The expensive frontier model is called once, and only ever on verified evidence — never on raw web pages.

01

Scope

The query is broken into targeted research tracks so the run stays focused, explainable, and free of wasted searches.

02

Gather

Hundreds of sources are searched, fetched, and reduced to atomic evidence cards — each fact bound to an exact source span for traceability.

03

Synthesise

The frontier model is called once, and only sees verified evidence cards — never raw pages. That keeps the reasoning layer auditable and the expensive work contained.

04

Verify

Every claim is checked against its source span so unsupported assertions are flagged and every conclusion keeps a clear evidence trail.

98%

Extraction Success

1000+

URLs per Job

99.4%

Verified Citations

4 min

Avg. Turnaround

Benchmarks

Analyst-grade output.
Auditable evidence trail.

A report-first system optimises for readable prose. Deep Water optimises for evidence you can trace. The difference shows up on every dimension that matters.

Dimension Deep Water Typical deep research
Sources consulted Hundreds to thousands A handful
Primary sources Papers, filings, specs General web
Verification Claim-to-span checked None
Evidence cards Structured output Prose only
Conflicting claims Surfaced with sources Hidden or averaged

And here's what that looks like on three real queries, run side by side.

Research query

“Latest approved treatments for triple-negative breast cancer”

Deep Water
Sources consulted 211 sources
Peer-reviewed papers 84 papers
Claim verification Every claim cited
Output Report + evidence cards
Typical AI deep research
Sources consulted ~12 sources
Peer-reviewed papers General web only
Claim verification Some claims cited
Output Single text block

Research query

“Compare REST vs GraphQL vs gRPC for high-throughput microservices”

Deep Water
Sources consulted 167 sources
Benchmarks / specs cited 31 benchmark refs
API & SDK docs included Yes, directly fetched
Output Structured comparison report
Typical AI deep research
Sources consulted ~8 sources
Benchmarks / specs cited General descriptions
API & SDK docs included Not directly
Output Prose summary

Research query

“Market size and growth drivers for solid-state batteries 2024–2030”

Deep Water
Sources consulted 118 sources
Market reports / filings 22 primary sources
Conflicting data flagged Yes, with source refs
Output Report + evidence cards
Typical AI deep research
Sources consulted ~10 sources
Market reports / filings Aggregated summaries
Conflicting data flagged Not typically
Output Single text block

For AI Product Builders

Built for teams shipping AI products.

Stop building research infrastructure yourself. Wire one engine into your product and ship the feature — instead of spending two quarters on scope, gather, extract, and verify.

Agent memory enrichment

Feed your agents verified, sourced evidence instead of raw search dumps. Memory your agent can cite — and you can audit.

Competitive intelligence

Continuous, structured tracking of competitors, pricing, and positioning — delivered as evidence your product can render or act on.

Compliance workflows

Outputs ship with a confidence-scored evidence trail, so regulated workflows have something concrete to review and retain.

Due diligence

Process hundreds of primary sources per target in minutes, with conflicting claims surfaced rather than averaged away.

Market intelligence

Market sizing, growth drivers, and segment analysis assembled from filings and primary reports — every number traceable to its source.

Ship it this sprint.

REST API, webhook callbacks, MCP, and an embedded runtime. One integration, and research becomes a backend you call — not a system you maintain.

Try for free

Research-Heavy Teams

Where verifiable research compounds

Beyond AI products, any team whose decisions depend on a defensible evidence trail — not just a readable summary — runs on Deep Water.

#1 Highest ROI

Growth & Product Teams

Series A–C startups, 5–10 person teams

Running 50–200 deep research queries a month — competitor analysis, market sizing, user sentiment. Currently burning $5K–20K/mo in analyst hours or hitting ChatGPT Pro caps.

You replace

2–5 analyst seats + $200/seat ChatGPT Pro

Top willingness-to-pay

VC & PE Analysts

Small funds, 3–5 analysts, deals on 48-hour cycles

300–800 due-diligence queries per month. Speed is the edge: your analyst processes 20 URLs in 3 hours, Deep Water processes 500 in 4 minutes. Incognito mode keeps deal flow private.

You replace

60% of junior analyst research time

Margin multiplier

Boutique Consulting & Market Research

Agencies billing $5K–50K per engagement

Cut per-report research cost by 80–90%. Deep Water produces a sourced first draft — your analysts refine and client-face. Same billable hours, higher margin.

You replace

$500–2,000 research cost per report

Land & expand

Enterprise Strategy Teams

Fortune 1000 market intelligence groups

Gartner and Forrester take 3–4 weeks and $100K–500K/year. Deep Water delivers the first-pass competitive landscape in 4 minutes. Use them together, not instead.

You replace

3-week research cycles

Developer-first

AI Product Builders

Founders shipping research features inside their own apps

Need a deep-research backend you can call from your own product? Forget building your own scope/gather/verify pipeline. REST API, webhook callbacks, bring-your-own keys in minutes.

You replace

3–6 months of infra engineering

Time saver

Literature Review & Policy Research

Researchers, think tanks, policy analysts

A 40-hour literature review takes 4. Triage 500 sources, find the 20 that matter, then read deeply. Every claim cited so you can walk the bibliography yourself.

You replace

Weeks of manual search & synthesis

The Foundation

We didn't build another agent.
We built the foundation underneath them.

Deep Water is the research layer that agents call when they need information they can trust.

API

REST + webhooks

CLI

terminal + scripts

MCP

agent-native

Embedded

in-process runtime

CLI — for individuals

Best for personal use

The fastest way to run deep research from your terminal. Zero install ceremony — npm i -g @unlikeotherai/deepwater, paste your key, done. Great for ad-hoc questions, side projects, research before a purchase, editor integrations, cron-driven briefings.

# install once
$ npm i -g @unlikeotherai/deepwater
# authenticate
$ deepwater config --set api-key=dw_xxx
# run a job
$ deepwater research "top 5 vector DBs"
  • One command, markdown on stdout
  • Credential stored locally, mode 600
  • 9 KB, zero runtime deps
  • Same $0.10 minimum. Same hard caps.
Install from npm
For teams & heavy users

Hosted API — for enterprise

Best for production & heavy volume

The same engine, exposed as a REST API built for scale. Webhooks instead of polling, per-project billing, bring-your-own-model-keys, higher rate limits, SLA. This is the path if research is a product feature, a backend workflow, or a recurring internal process.

# POST /v1/research
curl -X POST https://api.deepwater.live/v1/research \
  -H "authorization: Bearer dw_xxx" \
  -H "content-type: application/json" \
  -d '{"query":"...", "depth":"deep"}'
  • Webhook completion — no polling
  • Per-project keys, per-project billing
  • Bring your own model keys (BYO)
  • Higher rate limits & SLA on request
Try for free

Both paths hit the same engine, the same pricing, the same cost caps. The only difference is how the work enters and how the result comes out.

Pricing

Central tariff controls.
Immutable metering.

Every provider call is recorded with exact usage, cost, and attribution. Your centrally managed workspace plan owns commercial rating, statements, subscriptions, invoices, credits, and collection.

1 · Metering

Provider metering

Original provider counts, cost, and product attribution remain immutable.

2 · Plan

Workspace tariff

Default, organization, and team terms are assigned in one control plane.

3 · Statement

Commercial statement

One billing record contains rated usage, added value, subscription lines, and totals.

4 · Payment

Payment lifecycle

Checkout, invoices, credits, Portal, and collection stay centralized.

Workspace-specific

A default, free, at-cost, or custom tariff can be assigned at organization or team level.

Team visibility

Members see their own usage. Team owners and admins can review the same breakdown per user.

One billing record

Every call is metered once; the central billing authority applies commercial terms and owns the customer record.

See your exact tariff

Open Billing to see current plan configuration and immutable provider metering. Deep Water never derives a customer total locally.

View plan and metering

Enterprise

Custom

Custom monthly and usage terms, higher rate limits, an SLA, and embedded deployment for teams putting research on a critical path.

Talk to us
Launch offer · Limited

Research for the public good — at cost.

If you're running non-commercial research at a medical or pharma university, an ecological programme, or any initiative aimed squarely at human benefit, we'll run your jobs at cost during launch. No margin. No upsell. Just infrastructure you couldn't otherwise afford.

Who qualifies

Medical and pharma university research, ecological and climate studies, public-health work, and other non-profit programmes with a clear human-benefit thesis. Must be non-commercial.

How to apply

Reach out with a short description of the programme, the institution, and what you're trying to answer. We review everything individually and assign custom terms — there's no self-serve path for at-cost access.

Temporary, for launch

This is a launch-only programme. If it works we'll bring it back in later cohorts, but we can't commit to it indefinitely. If you need predictable access, apply now.

Apply for at-cost access

research@deepwater.live · Individual review · Terms assigned on approval

FAQ

Questions, answered.

Why not Gemini Deep Research?
Gemini generates a report for a human to read. Deep Water exposes research as infrastructure — its output is structured evidence that software, agents, and workflows can consume, gate, and act on programmatically.
Why not use OpenAI Deep Research?
Deep Water gives you API access, cost control, full visibility into the research process, and structured evidence as the output — not just a finished block of prose. It's built to be wired into products, not chatted with.
What makes Deep Water different?
Evidence-first architecture. The report is assembled from verified evidence rather than generated directly from web content. The frontier model only ever sees evidence that has already been checked against its source.
What are Evidence Cards?
Atomic facts, each linked to an exact source span, with a citation and a confidence score. They are the building blocks the report is assembled from — the unit of truth the whole pipeline operates on.
How are Evidence Cards different from citations?
A citation explains, after the fact, where a statement came from. An Evidence Card comes first — it's the verified building block the statement is constructed from. Citations annotate text; Evidence Cards produce it.
Can agents use Deep Water autonomously?
Yes. Deep Water is MCP-compatible and designed for agent workflows. An agent can commission research, read the verified evidence, act on it, and commission follow-up research — with no human in the loop.
How many sources can Deep Water analyze?
Depending on the depth you choose, anywhere from dozens up to thousands of URLs per job. You control how deep a run goes — and you only pay for the compute it uses.
Can I run Deep Water privately?
Public archive by default — that's what keeps public research at cost. A private archive is available if you need your jobs kept off the public archive, and incognito mode runs a job without storing anything: the report is delivered once, then deleted.
Why is Deep Water cheaper?
Cost is a consequence of the architecture, not a discount. The bulk of the work — reading and extracting evidence — is done by inexpensive models. The expensive frontier model is used only for the final synthesis, on a small, verified evidence set.
What happens when sources disagree?
Conflicts are surfaced, not hidden. Rather than silently averaging or picking one side, the report presents the conflicting evidence and the confidence attached to each — so you decide with the disagreement in view.

Stop generating research.
Start verifying it.

Give your agents and your team an auditable evidence trail instead of plausible text. One engine, exposed as API, CLI, MCP, and an embedded runtime.

Evidence-first architecture · Pay for compute, not seats · MCP-compatible