Every claim traced to its primary source
The Claim Verifier fact-checks every claim to its primary source, digs past the blogs that cite other blogs, and marks what it cannot verify rather than passing it through.
Verify · “market grew 30% last year”
Claim VerifierA growth stat to verify before it goes in the report. Tracing it to the source.
Several articles cite it. Following them back to see who did the original counting.
Found the primary source. The 30% figure checks out, for the segment defined the same way we mean it.
Logging the claim, verified to primary source1s
How it works
Claim to verify“is this stat real?”Draft submitted for checkReport before publish
Claim verified to sourceEcho chain discardedCorrected · 70% → 58%Unverifiable claim flaggedTriggered by a claim to check
A stat in a draft, a number in a report, or a claim pasted into Slack starts a verification. The agent takes it back to the source.
Traces past the echoes to the origin
It follows the citation chain to the primary source, discards the blogs that cite each other, and weighs how credible each source really is.
Logs each claim with its source
Verified claims land with a link to the primary source, corrected numbers replace the echoes, and anything unverifiable is marked, not passed through.
Without Relevance
With Relevance
A claim gets a quick search, and the first confident-looking source is taken as truth.
The agent traces each claim to its primary source before it accepts the number.
A stat from a blog citing a blog gets repeated, and the error travels with it.
The agent follows the chain to the origin, and discards the echoes citing each other.
A press release and a peer-reviewed study count the same in a rushed check.
The agent weighs how credible each source is and trusts the primary one over the loudest one.
Only the headline claims get checked; the supporting stats slide through unverified.
Every claim in the document gets traced, not just the ones that draw attention.
Rigor depends on who is checking and how much time they have.
One verification standard applied to every claim, regardless of who or when.
An unverifiable claim gets passed through so the report stays clean.
A claim with no primary source is marked unverified, never presented as confirmed.
Held to a quality bar, on every run
Every verification is checked against eval test cases written for this exact job before it is logged. If a run fails the bar, it never ships.
Claim VerifierSampling 2% of live runs · 97% passing this week
Shaped by your team, not a template
Experts and engineers tune Claim Verifier to your playbooks. Every change ships through the same evals.
Drag and drop
Compose agents and workforces on a visual canvas. Drag in agents, tools, and approvals in a flow anyone on the team can read.
Drag and drop
Compose agents and workforces on a visual canvas. Drag in agents, tools, and approvals in a flow anyone on the team can read.
Ticket TriagerBuild with AI
Describe the agent in plain language and Invent builds it: the prompt, the tools, and the evals to prove it works.
Build with AI
Describe the agent in plain language and Invent builds it: the prompt, the tools, and the evals to prove it works.
Claude Code
CodexBuild with MCP
Engineers drive the same platform from Claude Code, Codex, or Cursor. Create agents, link knowledge, and run evals over MCP.
Build with MCP
Engineers drive the same platform from Claude Code, Codex, or Cursor. Create agents, link knowledge, and run evals over MCP.
Built for enterprise teams
Run agents on your real data with the access controls, audit trails, and residency guarantees enterprises require all built in.
Security & data privacy
- Data residency
- PII masking
- Audit logs
- No training on your data
Access & controls
- Role-based access control
- SSO / SAML
- Human-in-the-loop approvals
- Version control
Monitoring & oversight
- Real-time monitoring
- Full agent tracing
- Cost visibility
- OTEL & Delta Share export
"The ability to be vendor agnostic and the ability to scale across a breadth of functions is a really key feature."
Levi Watters
Partner, KPMG Australia
"The key for us was how we can modularize industry knowledge and the best playbooks, and apply it."
Allen Roh
Senior Marketing Manager, Autodesk
"We're looking for every place where AI can allow sellers and customer success reps to be more engaged with customers."
Rob Giglio
Chief Customer Officer, Canva
"The ability to be vendor agnostic and the ability to scale across a breadth of functions is a really key feature."
Levi Watters
Partner, KPMG Australia
Team it up with more research agents
Market briefMarket Researcher
Researches a market from live sources into a sourced brief.
Source Synthesizer
Reduces many sources to cited findings, contradictions and all.
Insight · lean into BInsight Analyst
Finds the one insight in the data that changes a decision.
Themes written upInterview Synthesizer
Turns customer interviews into themes with verbatim evidence.
Every agent you add shares one stack
These agents all run on one platform: one gateway, one router, one eval suite, one audit trail. Every agent after the first ships faster.
- Triggered by key events or signalslike Zapier
- Given access to contextlike Zep
- Connected to your appslike Composio
- Access to all LLMslike OpenRouter
- Performance evaluatedlike Braintrust
- No-code agent builderlike Dust.tt
- Coordinated into teamslike CrewAI
- Kept alive through long-running worklike Temporal
- Traced at every step of every runlike Langfuse
Put the Claim Verifier to work this quarter
We partner with enterprise teams to take this work off their plate. An agent built with you, proven against your quality bar, and live in weeks.