Skip to content
Competitive Intelligence

The competitor brief, done by Monday morning.

A five-agent swarm reads the web for you and returns a sourced competitive intelligence report — positioning, funding, signals, and risks — in minutes. Built for VC associates and strategy leads who have to be right.

No credit card. Free 50-credit grant on signup. One CI run averages ~25 credits.

Used by analysts at
Why teams switch

Three problems that disappear.

The problem

Three browser tabs and a Slack ping at 11pm.

Junior analysts spend half a partner meeting cycle digging through Crunchbase, LinkedIn, news, and PR. The output is uneven and rarely sourced.

With Zyphv

A structured CI report, sourced and timestamped.

Every claim links back to a citation. Sections are consistent across runs. The associate spends the saved hours on judgement, not collection.

The problem

One-off prompts give one-off answers.

A single LLM call summarises, but it does not investigate. It cannot orchestrate research, cross-check signals, or compile a polished section.

With Zyphv

Five specialised agents working together.

Scout, Researcher, Mapper, Analyst, and Compiler each do one thing well. The Compiler hands you the brief — not a transcript.

The problem

You cannot defend a number you cannot trace.

IC memos that quote AI without a chain of evidence are unusable. Compliance, partners, and LPs will rightly ask where it came from.

With Zyphv

Provenance on every section.

Cryptographic hash of input and output per agent, plus source links. Verify what was produced, by which model, when. It is auditable.

How it works

Five agents. One brief.

Each step is sequential, sourced, and bounded. You can read the trace of every run.

01
Scout
Fast scan: company facts, key people, recent news.
grok-3-fast · web search
02
Researcher
Deep research with live search across funding, hiring, product signals.
grok-3 · live search
03
Mapper
Competitive landscape — adjacent players, segments, positioning.
claude-sonnet
04
Analyst
Strategic analysis, threats, opportunities, watch-list signals.
claude-sonnet
05
Compiler
Polished executive summary plus structured sections with citations.
claude-opus
Sample output

What you actually receive.

Rendered in-app and exportable to PDF. Below is a representative excerpt.

Competitive Intelligence — Acme Robotics
Run 8a3f · 5 agents · 24 credits · 7 sources
2026-06-04

Acme is moving downmarket; logistics incumbents are next.

Acme’s Q2 hiring shift toward warehouse integrations and a recent partnership with a 3PL suggest a near-term wedge into mid-market logistics — historically held by two slower incumbents.

$42M
Series B (Mar 2026)
+47
Logistics hires Q2
3
Named competitors

Premium-end automation has been Acme’s defensible niche. New job posts and the 3PL partnership indicate a planned push into mid-market — a deliberate adjacency, not a pivot.

crunchbase.comacme.com/bloglinkedin.comtechcrunch.com3pl-press-release
Pricing

Pay for credits. Use them on any swarm.

One CI run averages around 25 credits. Sign up to receive 50 credits, then top up as needed.

Starter
$9 / 100 credits

~4 CI runs. Best for evaluating quality on real targets.

  • All five CI agents
  • PDF + web report
  • Provenance + citations
  • Email support
Choose Starter
Scale
$69 / 1,000 credits

~40 CI runs. For diligence pods and corp dev teams.

  • Everything in Professional
  • Cross-run intelligence
  • API access (developer)
  • Slack / shared workspace
Choose Scale

Average CI run: ~25 credits. Cost depends on sources and focus areas.

We don’t promise magic. We deliver clarity when it matters most.

Questions

Reasonable answers.

How long does a CI run take?
Most CI runs finish in 4–8 minutes. You can watch progress live or come back later — the report waits for you.
What sources do the agents use?
Public web via live search, plus any files you upload (decks, prior memos, transcripts). You can toggle source categories per run.
Can I trust the citations?
Every section carries source links. Each agent action records a cryptographic hash of its input and output, so you can verify the chain end to end.
Can I export to PDF?
Yes. Every run renders a polished PDF with sections, callouts, and citations. You can also share via a tokenised link.
What if a run is bad?
If a run fails or returns an empty report, credits are refunded automatically. If you’re unhappy with quality, email us — we read every message.
Do you train on my data?
No. Uploaded files are tenant-isolated and never used to train models. We use API providers under their no-training terms.
More from Zyphv

Explore other intelligence verticals.

Ready when you are.

Start free — 50 credits on us. No card. No noise.

Begin your first CI run