Talk to your data in Claude, ChatGPT, Codex and more.
Learn more

Your marketing agents
need a context platform.

Everything your systems record, and what every number means. Served to any agent through a single governed MCP.

Get started for freeRequest a demo
Try a question from
RevenueMarketingCustomer success
Claude
How much new MRR did we add in September, and from whom?
Used
Nekt
3 steps
Read 2 context documents
Revenue & MRR, Active customers
Ran
get_relevant_tables_ddl
Ran
execute_sql
on
trusted.mrr_daily
1.4 s

You added $42,180 in new MRR in September, 19% more than August1. Expansion drove 61% of it, led by Apex Studios, Marlin Labs, and Northwind3.

New MRR by month
USD, 2026

Two internal test workspaces are left out, per your Active customers definition2.

Sources
1
trusted.mrr_daily
Daily MRR by account and plan, in USD. Runs every hour.
Table from your catalog
2
Active customers rules
Paying means active, or past due for under 14 days. Internal workspaces never count.
Document from your context graph
3
Revenue & MRR
New MRR counts new logos and expansion, converted at the month's closing rate.
Document from your context graph
ChatGPT
Which channels drove the most conversions this quarter?
Used
Nekt
3 steps
Read 1 context document
Sales funnel stages
Ran
get_relevant_tables_ddl
Ran
execute_sql
on
trusted.campaign_performance
0.9 s

Google Ads leads with 1,248 conversions, 34% of the quarter1. LinkedIn Ads is second with 876, and its leads reach SQL at the highest rate1.

Conversions by channel
Q3 2026

Partner referrals are left out, per your funnel stage definitions2.

Sources
1
trusted.campaign_performance
Conversions, spend, and pipeline by channel and campaign. Runs every hour.
Table from your catalog
2
Sales funnel stages
What counts as an MQL, SAL, and SQL. Partner referrals never count as marketing leads.
Document from your context graph
Codex
Which accounts are at risk of churning this month?
Used
Nekt
3 steps
Read 2 context documents
Product usage, Active customers
Ran
get_relevant_tables_ddl
Ran
execute_sql
on
trusted.account_health
1.1 s

3 accounts are at risk, worth $18,420 in MRR, 3.8% of the total1. Each used less than 30% of its plan over the last 30 days, your churn-risk signal2.

At-risk accounts
Last 30 days
Account
MRR
Plan usage
Silent
Beacon Group
$8,200
22%
21 days
Harbor & Co.
$6,020
18%
12 days
Lumen Labs
$4,200
27%
9 days

Only paying accounts are included, so trials and internal workspaces don't skew the list3.

Sources
1
trusted.account_health
One row per paying account, with MRR and 30-day product usage.
Table from your catalog
2
Product usage
Under 30% of plan usage for 30 days is the churn-risk signal.
Document from your context graph
3
Active customers rules
Paying means active, or past due for under 14 days.
Document from your context graph
Same context in
Claude
ChatGPT
Codex
Cursor
VS Code
Lovable
and any MCP client

Trusted by growing companies worldwide

Make your company queryable.

To take action, agents call tools. To answer questions, they need annotated columns, queryable history, and one source of truth. The data platform delivers all three.

Answers with tool calls
▸ stripe.list_payment_intents
ok
↳ paginating… 174 customers
▸ stripe.list_payment_intents
ok
▸ stripe.list_payment_intents
ok
… 17 more pages …
▸ hubspot.search_companies
429
⚠ rate limited — retry in 2s…
▸ hubspot.search_deals
ok
↻ disambiguating duplicates · matching by stripe_customer_id
▸ hubspot.search_deals
…
… still fetching at 14m…

Agent: "Best-effort answer. 234 tool calls,
        14 minutes, partial confidence."

234
tool calls
83%
accuracy
15m
latency
234K
tokens
Answers with Nekt
SELECT name,
       revenue,
       stage,
       status
FROM nekt_trusted.customers
ORDER BY revenue DESC
LIMIT 20;
Name
Revenue
Stage
Status
Apex Studios
$70,000
> 12 mo
Open
Marlin Labs
$19,977
Deal won
Won
Northwind Inc.
$11,976
> 12 mo
Open
Beacon Group
$11,976
Deal won
Won
1
sql query
100%
accuracy
19s
latency
18K
tokens
Direct tool calls
With Nekt
▸ stripe.list_payment_intents
ok
↳ paginating… 174 customers
▸ stripe.list_payment_intents
ok
▸ stripe.list_payment_intents
ok
… 17 more pages …
▸ hubspot.search_companies
429
⚠ rate limited — retry in 2s…
▸ hubspot.search_deals
ok
↻ disambiguating duplicates · matching by stripe_customer_id
▸ hubspot.search_deals
…
… still fetching at 14m…

Agent: "Best-effort answer. 234 tool calls,         14 minutes, partial confidence."

234
tool calls
83%
accuracy
234K
tokens
SELECT name,
       revenue,
       stage,
       status
FROM nekt_trusted.customers
ORDER BY revenue DESC
LIMIT 20;
name
revenue
stage
status
Apex Studios
$70,000
> 12 mo
Open
Marlin Labs
$19,977
Deal won
Won
Northwind Inc.
$11,976
> 12 mo
Open
Beacon Group
$11,976
Deal won
Won
1
sql query
100%
accuracy
18K
tokens
Agents connected directly to APIs
38%
Agents + Nekt
91%
Median across 5 runs of Anthropic Claude Opus 4.7. Same gap shows on Sonnet 4.6 (121× cheaper, +34pp accuracy) and Haiku 4.5 (54× cheaper, +73pp accuracy).

Connect. Govern. Serve.

Three layers of the data platform, built as a single system.

1
2
3

Connect

+200 connectors — SaaS, databases, files, MCP. Nekt handles the sync, the schema, and everything in between.

Shopify
Stripe
Postgres
GA4
+200
See connectors

Govern

Transform, annotate, and govern — SQL, notebooks, semantic layers, permissions, historical, and lineage, all built in.

Raw204
Trusted81
Service37
Explore the layers

Serve

MCP server, Data API, and native destinations — every agent, every dashboard, and every teammate gets the same trusted answer.

Claude
Codex
Lovable
PowerBI
See integrations
Connect

Connect every source your business run on.

Works with SaaS apps, databases, webhooks, and flat files (CSV, Excel, Parquet), plus unstructured data like PDFs, images, and video. Ready for every agent, dashboard, and team.

Explore our 200+ connectors
Nekt Storage

One governed source of truth, built by Nekt.

Structure your own data layers. Annotate every column so agents query with context. Hosted on Nekt Cloud, your GCP, or your AWS.

Self-serve access

Any team member can explore data without waiting on engineering.

Always current

Incremental syncs keep every table fresh, no stale dashboards.

Right access, right people

Permissions at the table and layer level keep data safe.

All your data, connected

Orders, customers, and more, linked in one place, not siloed per tool.

One governed source of truth, built by Nekt.

Structure your own data layers. Annotate every column so agents query with context.
Hosted on Nekt Cloud, your GCP, or your AWS.

Search

Raw

76

Amplitude

6

Hubspot

11

Facebook Ads

17

Financial

7

Intercom

3

LinkedIn Ads

12

Stripe

4

Product Images

Trusted

11

Sales

5

Customer Service

6

Service

12

Orbit's Catalog

/

Raw

/

Facebook Ads

Docs

Help

Facebook Ads

Data from Nekt's Facebook Ads account.

Search

Tags
Marketing
Permissions
A
All

Name

Description

Created at

Updated at

facebook_ads_images

Image assets uploaded to the ad account, with hashes, dimensions, and linked creatives.

Image assets uploaded to the ad account, with hashes, dimensions, and linked creatives.
Description enriched by Analyst

Oct 7, 2025

Mar 1, 2026

facebook_ads_metrics

Core delivery metrics — impressions, reach, frequency, clicks, CTR, CPC, and CPM.

Core delivery metrics — impressions, reach, frequency, clicks, CTR, CPC, and CPM.
Description enriched by Analyst

Oct 7, 2025

Mar 1, 2026

facebook_ads_campaigns

Campaigns with objective, status, budgets, bid strategy, and special ad categories.

Campaigns with objective, status, budgets, bid strategy, and special ad categories.
Description enriched by Analyst

Oct 7, 2025

Mar 1, 2026

facebook_ads_audiences

Custom and lookalike audiences, with size estimates, source, and retention settings.

Custom and lookalike audiences, with size estimates, source, and retention settings.
Description enriched by Analyst

Oct 7, 2025

Mar 1, 2026

facebook_ads_performance

Spend, ROAS, quality rankings, and cost-per-action across campaigns, ad sets, and ads.

Spend, ROAS, quality rankings, and cost-per-action across campaigns, ad sets, and ads.
Description enriched by Analyst

Oct 7, 2025

Mar 1, 2026

facebook_ads_adsets

Ad sets with targeting, schedule, budget, optimization goal, and learning phase status.

Ad sets with targeting, schedule, budget, optimization goal, and learning phase status.
Description enriched by Analyst

Oct 7, 2025

Mar 1, 2026

facebook_ads_insights

Daily breakdowns by age, gender, country, device, placement, and region.

Daily breakdowns by age, gender, country, device, placement, and region.

Description enriched by Analyst

Oct 7, 2025

Mar 1, 2026

facebook_ads_conversions

Custom conversions tracking URL and event-based actions, with last-fired timestamps.

Custom conversions tracking URL and event-based actions, with last-fired timestamps.

Description enriched by Analyst

Oct 7, 2025

Mar 1, 2026

Raw in, business-ready out.

Shape your data layers with Queries or Notebooks in Python, PySpark, or web scraping. Every pipeline scheduled, versioned, and production-ready.

Raw
raw.stripe.subscriptions
+ 2 sources
email
plan
mrr
Ann@acme.io
Pro-Annual
2400
ANN@acme.io
pro annual
2400
jk@marlin.com
null
149
rc@apex.com
STARTER
89
37% duplicates
12 plan spellings
Missing values
SQLPython
SELECTLOWER(email) email,
m.clean AS plan,
MAX(mrr) mrr
FROMraw.stripe.subscriptions
JOINtrusted.plan_map m
GROUP BY1, 2
df = nekt.read(
"raw.stripe.subscriptions")
df = df.withColumn("email",
lower("email"))
df = df.join(plan_map, "plan")
df.groupBy("email").max("mrr")
Trusted
trusted.account_health
ClaudeCodex
Ready for agents
email
plan
mrr
ann@acme.io
pro_annual
2400
jk@marlin.com
unknown
149
rc@apex.com
starter
89
md@northwind.io
enterprise
4800
plan
Normalized plan: starter, pro_annual, or enterprise

SQL, Python, or PySpark

Write it yourself, or let your agent draft it.

Scheduled and versioned

Runs on your schedule, with every change in GitHub.

History built in

Histories, lineage, and data expectations keep it honest.

Context Graph

Teach your agents how your business runs.

Tables hold the numbers, your team holds the definitions. Write down what counts as an active customer or how MRR is calculated once, and every agent follows it.

Let your agent interview you and draft the first documents

Scope which agents can read each document

With context graphWithout context graph
Claude
How many active customers do we have?
Read a context document from
Nekt
You have
1,212
active customers.

Excluded 64 trials and 8 internal workspaces, per your Active customers definition.

Active customers rules
Document from your context graph

Paying means Stripe status active, or past_due for under 14 days.

  • Trials never count
  • Internal workspaces never count
AS
Edited by Ana Souza
Claude
How many active customers do we have?
You have
1,284
active customers.

Counted every row in stripe.subscriptions, including trials and internal workspaces.

No definition to read, so the agent guesses.

Nekt delivers the data, agents do the work.

Build with AI on Nekt's Platform API, Data API, or MCP server. Claude, Lovable, ChatGPT, OpenClaw, or your own agent share the same memory layer: governed business data, annotated for context, fresh on the schedule you set.

Everyday at 3pm
Stripe
HubSpot
Postgres
Claude
You

Build me a RevOps daily briefing. Pull from my Stripe, HubSpot and Postgres data. I want MRR by plan, the accounts adding the most this month, and anyone at risk of churning.

briefings.company.com/revops-daily
RevOps Daily Briefing
Thursday, April 23
Total MRR
$487,210
8.3%
vs. $449,810 last month
New MRRApril
$42,180
19.2%
71% of $59k monthly target
14 net-new logos
Open pipeline
$312,400
27 deals
Discover → Demo → POC → Close
Churn risk
$18,420
3 accounts
3.8% of MRR
MRR by plan
Last 6 months
EnterpriseProStarter
$500k$400k$300k$200k$100k$0
Nov
$358k
Dec
$382k
Jan
$410k
Feb
$432k
Mar
$458k
Apr
$487k
Plan mix
This month, by MRR
$487k
Total MRR
Enterprise50%$245k
Pro Annual32%$156k
Starter6%$28k
Analyst

Your data has answers. Just ask.

Nekt Analyst turns natural language questions into SQL queries, executes them instantly, and delivers actionable insights — no technical expertise required.

Top acquisition channels
Monthly active users
Customer churn rate
Revenue by region
Analyst
f

Facebook Ads

Campaigns

S

Stripe

Payments

Which acquisition channels drove the most revenue last quarter?

Analyst

Here are your top acquisition channels for Q1 2026, ranked by revenue. I joined Facebook Ads campaign data with Stripe payments to attribute each customer to their first-touch source.

sql
Run

5 rows

2m ago

Save
Open on Explorer
channel
customers
revenue
roas
Facebook Ads
1,248
R$ 287,450
5.4x
Google Ads
892
R$ 198,320
7.9x
Instagram
643
R$ 156,890
4.1x
TikTok Ads
421
R$ 94,560
2.8x
Organic Search
387
R$ 88,210
22.1x
Expand table

Ask anything…

Manage context2
Try Analyst free
Integrations

Your data, connected to every AI tool.

Give Claude, ChatGPT, and any AI agent direct access to your data through Nekt's MCP Server and Data API.

Nekt MCP

The fastest way to connect AI to your data

The Model Context Protocol is the open standard for AI-to-data communication. Nekt speaks it natively — set up in under a minute.

Works with

Claude
Claude
Codex
Codex
Lovable
Lovable
OpenClaw
OpenClaw
and more
Share
How are our marketing campaigns performing this quarter?
Used Nekt integration, loaded tools›

The trusted.campaign_performance table tracks all campaign metrics by channel. Let me query it.

Used Nekt integration›

Here's a breakdown of marketing campaign performance this quarter:

Top channels by conversions:

RankChannelConversions
1Google Ads1,248
2LinkedIn Ads876
3Email Nurture641
4Organic Search539
5Meta Ads412

Key takeaways:

Google Ads leads with 1,248 conversions — 38% of total pipeline, indicating strong paid search ROI.

LinkedIn is a strong #2 with 876 conversions, making it the top channel for B2B lead quality.

Platform API

Execute SQL queries and discover schemas via REST. Build in any language or connect your BI tool of choice.

Data API

Manage sources, catalogs, and pipelines programmatically. Terraform provider included.

Governance

Data permissions for every individual, every agent, and every team.

Direct permissions for individuals. Scoped API tokens for agents. Group-based access for teams. Every rule enforced at the data layer, down to the table.

3 layers
1 folder
4 tables

Permission people down to the table.
Grant edit, view, or manage rights across entire layers, single folders, or specific tables.

FinancesGroup
12 members

Manage access by team, not by person.
Bundle people into groups like, then assign permissions once instead of user by user.

ClaudeAPI Token
i29dbm....FWsOP8

Give every agent its own scoped token.
Issue API tokens with granular permissions, so agents only touch what they should.

Turn data into action-ready insights.

Explore, transform, and visualize business data across every team to build a data-driven culture that accelerates growth and drives revenue impact.

Explorer

Ask smart questions and run analyses with SQL, Python, or AI.

Notebooks & Queries

Organize metrics and prepare trustworthy, recurring datasets.

Destinations

Sync trusted data from Nekt directly into your tools.

SOC 2 Type II compliant, with independently audited security controls.

As seen on

TerraUolPequenas Empresas & Grandes NegóciosExameThe ShiftStartups.com.br
Security

Built for teams that take security seriously.

Enterprise-grade protection at every layer. Independently audited, always encrypted, zero exceptions.

Storage

Deploy in your cloud or on Nekt Express with same security controls apply.

SOC 2

Type II certified, independently audited security controls, verified continuously.

Permissions

Role-based access gives you granular permissions down to the table level