New Wyoming + New Mexico now live →

Clean oil & gas data,
ready to forecast.

State agencies still ship well data in formats from 1965. CleanWell decodes EBCDIC, normalizes 5 states, and delivers ComboCurve-ready datasets — automatically, every morning.

See how it works

1,876,582

Wells in production

States covered

1.18M

Monthly production records

6:00 CT

Daily refresh

cleanwell.io / wells

Wells · 1,876,582

Updated 6:04am CT

WellAPI-14StateCum Oil (bbl)Status

SCHLITTLER 1H42503802050101TX186,420Active

HERMOSO 14-2H35025034220000NM472,103Active

RED HILL B-940103128440000OK28,917Inactive

PIONEER 24-A42389210770000TX912,488Active

BLACK MESA 1708123005740000CO5,221Plugged

SAGE CREEK 2-156019100330000WY63,440Active

The Problem

Public well data is stuck in the mainframe era.

Every operator, fund, and forecaster needs this data. It's public, it's free — and it's nearly unusable without weeks of work.

Legacy formats no one can read

Texas RRC publishes data in EBCDIC — 1960s IBM mainframe binary. Production numbers are stored as packed decimal (COMP-3). Excel chokes. Python chokes. Vendors charge $50K/yr to decode it for you.

Weeks of manual cleanup, every quarter

Teams burn 2–6 weeks per quarter downloading, decoding, and stitching together 12+ files. By the time the data lands, half of it is already stale.

Wells and production don't match up

Production is reported at the lease level. A single lease can have dozens of wells. Mapping production to a well requires cross-referencing three datasets and reconciling mismatched district codes.

Every state does it differently

Texas = EBCDIC. Oklahoma = CSV. New Mexico = SQL Server XML over FTP. Colorado = daily-rebuilt shapefile. There's no standard — and the work multiplies with every basin you cover.

olf008l.ebcEBCDIC · COMP-3

f1f0 f7c2 f0f0 f0f1 f4f0 f0f1  ................
f0f0 f0f0 f0f0 f0f0 f0f0 e6c9   ................
d3c4 c3c1 e340 4040 4040 4040   .@@@@@@@@@@@@@@@
4040 4040 4040 4040 4040 f0f0   @@@@@@@@@@......
f0f0 f4f9 f0f4 f9f0 f5f9 f0f8   ................
f3f0 f9f3 f1f9 f7f2 f0f2 f0f3   ................
0000 0000 7563 6c40 4040 4040   ........ucl@@@@@

1,200 bytes per record · no headers
no delimiters · binary packed decimals

dbf900.txtASCII · Fixed-width 247B

0100100001010106001 19631027000
000000000000000000 000000000010
000100000000000000  NNN00000000
Y0 13A199309000000199310221996
02L0040230000000199801000000Y

Signed numerics use "overpunch"
encoding on the last byte.

Before / After

From unreadable to forecast-ready.

Real bytes from Texas RRC, real JSON from CleanWell. The "after" panels are the exact format ComboCurve's API accepts — no reformatting on your end.

Before · Raw Wellbore Record247-byte fixed-width

01 001 00001 01 01 06 001 19631027 000000000000000000000 0000000000 10000100000000000000 NNN0000000 0Y0 13A19930900000019931022 1996 02L00402300000001998 01000000Y Field district: 01 API county: 001 API unique: 00001 Compl. date: 19631027 (raw string) Latitude: ??? (overpunch encoded) Longitude: ??? (overpunch encoded) District codes: don't match oil master

After · CleanWell Well HeaderComboCurve-ready JSON

{ "dataSource": "rrc_texas", "api14": "42503802050101", "api10": "4250380205", "name": "SCHLITTLER", "latitude": 33.064692, "longitude": -98.948535, "wellStatus": "ACTIVE", "trueVerticalDepth": 3124, "completionDate": "1984-01-12", "county": "YOUNG" } // API-14 built from FIPS + county + unique // Lat/long decoded from signed overpunch // District codes normalized across datasets

Before · Lease-level ProductionEBCDIC COMP-3 packed

Offset 0x000: f1f0 f7c2 f0f0 f0f1 f4f0 Offset 0x00C: f0f0 f0f0 f0f0 f0f0 f0f0 COMP-3 packed decimal example: Bytes: 0x00 0x12 0x34 0x5C Decode: nibbles 0,0,1,2,3,4 sign=C(+) Result: +1234 Production reported per LEASE, not well. Lease 00014 in District 7B has 15 wells — one aggregate number for all of them. 12 monthly volumes repeat in a single 1,200-byte record. No API number here.

After · Monthly ProductionComboCurve-ready JSON

{ "Date": "2025-12-01", "Oil": 258, // BBL "Gas": 5364, // MCF "Water": null, "DaysOn": 31, "ChosenID": "42151332460000", "DataSource": "rrc_texas" } // Lease production resolved to well API-14 // COMP-3 packed decimals decoded // CCYYMM dates parsed to ISO 8601

How it works

Three steps. Zero mainframe expertise required.

Agencies

TX RRC, OK OCC, CO ECMC, WY WOGCC, NM OCD

Decode

EBCDIC, COMP-3, overpunch, shapefile, XML

Transform

dbt models on Snowflake. Wells linked to production.

Deliver

Web UI · CSV · REST API · ComboCurve · Snowflake share

We pull the raw data daily

Dagster runs at 6:00 AM CT, automatically pulls the latest files from every state agency, and checks freshness before reprocessing.

Texas RRC (EBCDIC + ASCII)
Oklahoma OCC (RBDMS CSV)
Colorado ECMC (daily shapefile)
Wyoming WOGCC (WFS)
New Mexico OCD (FTP XML)

We decode, clean, and link

Our pipeline does the work that normally takes weeks of an engineer's time — every night, deterministically.

EBCDIC & COMP-3 decoding
Signed overpunch numeric parsing
API-14 number construction
Lease → well production mapping
District code normalization

You get ComboCurve-ready data

Well headers and monthly production formatted exactly how ComboCurve expects them. Pull via UI, API, or live Snowflake share.

ComboCurve well upsert schema
ComboCurve MonthlyProductionInput
CSV, JSON, REST API
Snowflake Marketplace listing
Auto-push into your CC account

What you get

Everything you need to forecast a well.

1.87M wells, 5 states

Every well ever drilled in TX, OK, CO, WY, and NM — with operator, completion date, depth, type, and coordinates.

TX · OK · CO · WY · NM

Full production history

1.18M monthly oil (BBL), gas (MCF), and days-on records. Texas lease-level data resolved to well level by our linking engine.

→ 656K ComboCurve-ready rows

GPS coordinates, WGS84

Latitude and longitude for every well with location on file. West longitudes corrected. Signed overpunch decoded to proper decimals.

ComboCurve-native format

Output matches ComboCurve's API schema exactly — well upsert and MonthlyProductionInput. Zero reformatting, zero field mapping.

Snowflake data share

Premium tier: live, zero-copy access in your own Snowflake account. We update — you query. Optionally list on the Snowflake Marketplace.

Daily refresh, monitored

Pipeline runs daily at 6:00 AM CT. Staleness detection only reprocesses changed files. You see freshness on every row.

Who it's for

Built for the teams who actually use the data.

E&P operators

Forecast your wells and offset wells without waiting weeks on a data vendor's quarterly drop.

Reservoir engineers

Production histories linked to well identifiers — ready for decline-curve analysis in ComboCurve or your own notebook.

Land & title

Look up any well by API across five states. Completion dates, operator, field, and lease attached.

PE & energy funds

Due diligence with clean production data across entire basins. No vendor contract required.

ComboCurve users

Skip the manual import. We push directly into your CC account in the exact schema it expects.

A&D advisors

Build comp sets and valuation models faster across TX, NM, OK, CO, WY.

Pricing

Simple plans. No vendor lock-in.

Start with what you need. Scale when you're ready.

Essential

$500 / month

Well identifiers and production history for a single state.

API-14 well identifiers
Full monthly production
Single state of choice
CSV & JSON exports
Monthly data refresh

Stop wrangling. Start forecasting.

Join teams who replaced their quarterly data wrestling with a single morning sync.

Wells

1,876,582

↑ 4,201 this week

Production records

1,180,836

↑ 28,194 this week

States covered

TX · OK · CO · WY · NM

Latest production

2026-04

RRC · last month

Basin

Play

Formation

Well type

RRC district (TX)

Completion date

TVD (ft)

Cum oil (bbl)

No advanced filters applied.

Wells1,876,582 · Cum Oil8.4B bbl · Cum Gas21.7T mcf · Avg TVD9,432 ft Latest prod2026-04

	Well	API-14	State	Operator	Basin / Play	Status	Cum Oil (bbl)	Cum Gas (mcf)	First Prod

Showing 1–25 of 1,876,582

1 2 3 … 75,063

—

Identification

Location

Geology

Drilling & completion

Production summary

Cum oil

—

Cum gas

—

Cum water

—

Monthly production · last 24 months

Oil & gas

Oil (bbl) Gas (mcf) Water (bbl)

Production history · last 12 months

Date	Oil (bbl)	Gas (mcf)	Water (bbl)	Days on

ComboCurve payload