Have a question about a Texas business or property? Open the kit in an AI agent that can read files and run commands. Ask it to find the records, save the results, and explain what they show.
See the Cameron County example for a starter prompt, an actual agent walkthrough, and saved results. A free business lookup sample is also available.
A self-hosted toolkit with 10 verified Python scrapers for business records, Travis County property records, professional licenses, campaign finance, oil and gas data, and environmental violations. The included trust manifest records each supported scraper's validation date, query, and result.
The kit also includes 261 beta TypeScript modules on the bundled FeedEater runtime and a 764+ source catalog of inventoried government data sources. Catalog entries document sources; they do not represent 764 working scrapers. Beta modules are validated on a rolling basis. Check each source's status and requirements before relying on it.
One question, a live search, and a saved report. This walkthrough uses the released v1.0.2 ZIP with an actual Codex CLI session. It includes narration and captions.
Copy the prompt · Read the agent’s report · Download the records · Read the actual session transcript
Texas has data-center projects, energy infrastructure, industrial construction, and Starbase. For people researching development, that means questions about land, ownership, and the businesses involved.
We started with Texas to help people investigate those projects. The records are spread across state agencies and counties. The kit supplies commands for supported sources so your agent can start with a search instead of building each scraper.
Texas offers tax incentives for qualifying data centers, but a proposal still needs checking. The state ordered an audit of data-center grid-connection requests in August 2026. Project announcements, approvals, and completed purchases are different records.
Download and unzip the kit, then open the extracted folder in your agent’s workspace. In v1.0.2, START_HERE.md explains the first steps and AGENTS.md provides the agent instructions. Ask your agent to read AGENTS.md, check the source coverage, and set up the dependencies for your task. Older ZIPs use the README. It needs file access, permission to run commands, and internet access to query government sources.
Read AGENTS.md. Find Texas business records matching this company name, save the results, and summarize them with source links. Tell me if the source is unavailable or the kit cannot answer part of my question.
The kit provides Python commands and a REST API for supported sources. Before writing a new scraper, check AGENTS.md, docs/CLI_REFERENCE.md, and the source validation records for an existing match. In older ZIPs, start with the README. The optional TypeScript modules have their own requirements and validation status.
# Start the optional module runtime
$ docker compose --profile modules up -d
# Open the module web UI at http://localhost:3000
# Check the module API
$ curl localhost:4000/api/healthThe optional module runtime includes a web UI, API, job worker, PostgreSQL, Redis, and NATS. The Python scrapers also work independently of this runtime. Modules start disabled; enable only the sources you need. Browser-dependent beta modules require additional setup and are not supported in the stock container.
For supported download sources, retrieve published datasets. The verified RRC bulk scraper provides a catalog of 22 downloadable datasets; availability and file sizes vary.
Some beta modules collect records across multiple pages or identifiers. Coverage, rate limits, and resume support depend on the module; check its documentation and validation status.
Search supported sources with parameters such as a business name, address, or well API number. Accepted queries and result limits are documented per source.
With the module runtime running, open localhost:3000 to browse modules, edit settings, trigger jobs, and watch logs.
The module UI is separate from the Python CLI and REST API. A module appearing in the UI does not mean its source has been live-validated.
# List sources in the Python toolkit
$ curl localhost:8000/sources
# Check collection status
$ curl localhost:8000/status
# Read stored records
$ curl localhost:8000/recordsThese examples use the Python API on port 8000. The optional FeedEater module API runs separately on port 4000, with its web UI on port 3000. Consult the included documentation for each runtime.
Business entity lookup through the Comptroller API, including available officers, registered agents, and filing information.
Franchise tax account status and right-to-transact information. Broader Comptroller sources are documented separately in the catalog and beta modules.
The verified property scraper covers Travis County (Austin metro). The broader county catalog is not a claim of verified coverage across all 254 Texas counties.
Campaign contributions and expenditures, plus the filer registry for candidates and committees.
Well search, GIS well and pipeline layers, and a bulk-download catalog. Coverage and query limits are documented for each scraper.
Environmental violations and enforcement records, plus professional license searches. Other agency sources in the catalog are not part of this verified set.
| Runtimes | Python scrapers; optional FeedEater TypeScript modules |
| Messaging | NATS |
| Database | Python: SQLite or PostgreSQL; module runtime: PostgreSQL |
| Containerization | Docker + Docker Compose |
| API | Python REST API on port 8000; module API on port 4000 |
| Web UI | Lightweight dashboard at localhost:3000 |
| Verified scrapers | 10 Python scrapers with recorded live-validation evidence |
| Beta modules | 261 TypeScript modules; validation and requirements vary |
| Source catalog | 764+ inventoried sources; not a count of working integrations |
| Collection Modes | Bulk, Crawl, Query |
| Output Formats | JSON, CSV, PostgreSQL |
| Rate Limiting | Source-specific; consult each scraper or module |
| Retry Logic | Exponential backoff with configurable retries |
| Resume | Module-specific; consult the included documentation |
| Logging | Structured JSON logs, configurable verbosity |
| License | Source-available; see the included license. Converts to MIT 12 months after purchase |
| Updates | All v1.x updates included |
Is this legal?
Yes. All sources are public records under Texas law. This tool accesses the same data available on government websites.
What do the three counts mean?
10 Python scrapers have recorded live-validation evidence. The 261 TypeScript modules are a broader beta collection, validated on a rolling basis. The 764+ catalog entries document inventoried sources, including sources outside the supported scrapers and beta modules. These counts are not interchangeable.
What do I need to run it?
Python 3.11+ for the Python toolkit. The optional module runtime can be started with docker compose --profile modules up -d. Running modules without Docker requires Node 20+ and the documented backing services. Some beta modules also require a browser.
Can AI agents use this?
Yes. The Python CLI and REST API expose source information, queries, and results. Agents can use the trust manifest to distinguish verified, experimental, and unsupported sources. The optional FeedEater runtime has a separate module API.
What happens after 12 months?
Keep your downloaded copy. All v1.x updates are included through your Gumroad library. The code license converts to MIT 12 months after purchase.
Do I need an API key?
The 10 verified Python scrapers do not require paid data-provider API keys. They use government websites and public APIs directly. Check individual beta modules for additional requirements.
Can I modify the code?
You can modify the code. The included license sets the conditions for use and distribution during the first 12 months; it converts to MIT after that period.
How is this different from BatchData or PropStream?
This is a $13.37 one-time purchase of tools you run on your own machine. You get 10 verified scrapers, 261 beta modules, and a 764+ source catalog across several public-record categories. It is not a hosted database or a guarantee of comprehensive Texas coverage.
Will there be other states?
Texas is our first kit. The current source list covers Texas.
Download the kit and give your agent a Texas research task.
Buy on Gumroad for $13.37