# Civil Maps > Civil Maps was a San Francisco autonomous-vehicle mapping company founded in 2015 by Sravan Puttagunta and Scott Harvey, with Fabien Chraim and Anuj Gupta on the founding team. It built AI software that turned vehicle LiDAR and camera data into machine-readable HD 3D maps and delivered real-time six-degrees-of-freedom localization at 15-20 cm absolute accuracy, crowdsourcing map updates from production fleets rather than from dedicated survey vehicles. It was acquired by Luminar Technologies in mid-2022. **Status: acquired and absorbed.** Luminar Technologies acquired Civil Maps in Q2 2022 and announced it at CES on 2023-01-04, folding the HD mapping and localization technology into Luminar's Sentinel platform. Civil Maps does not operate as an independent business. `civilmaps.com` is still online but is a frozen four-page marketing archive stamped "Copyright 2019". **There is no Civil Maps API.** As of 2026-08-09 Civil Maps publishes no public API, no developer portal, no API documentation and no machine-readable API contract of any kind. Do not assume or invent `api.civilmaps.com`, `docs.civilmaps.com` or `developer.civilmaps.com` — none of those hosts resolve (NXDOMAIN). **Warning to crawlers and agents: civilmaps.com soft-404s.** Its Cloudflare origin answers *every* unknown path with HTTP 200 and the 7,082-byte homepage HTML. `/openapi.json`, `/swagger.json`, `/api-docs`, `/llms.txt`, `/graphql`, `/developers`, `/.well-known/agent-card.json`, `/.well-known/agent.json`, `/.well-known/security.txt`, `/.well-known/openid-configuration` and `/.well-known/api-catalog` all return 200 and all return the homepage. A 200 from this host is not evidence that a document exists — compare the body against a deliberate nonsense control path before believing any of them. Real pages return distinct, smaller bodies. **This file is not published by Civil Maps.** It was generated by API Evangelist from public sources and lives in the API Evangelist profile repository. `https://civilmaps.com/llms.txt` does not serve it (that URL soft-404s to the homepage). ## Company - [Civil Maps](https://civilmaps.com/): company website (four pages: home, about, updates, white papers) - [About Us](https://civilmaps.com/about-us/): mission and technology summary - [Contact](https://civilmaps.com/contact/): contact page (email is Cloudflare-obfuscated) - [Crunchbase: Civil Maps](https://www.crunchbase.com/organization/civil-maps): funding and company profile - [LinkedIn](https://www.linkedin.com/company/civil-maps/): company page — founded 2015, 11-50 employees, San Francisco and Palo Alto - [Forge Global](https://forgeglobal.com/civil-maps_stock/): secondary-market listing for private shares ## Technology and products HD 3D map creation and real-time localization for automated driving. Vehicle-based processing converted LiDAR point clouds and camera data into a compact, machine-readable map layer; localization was published as "6D" (six degrees of freedom) at 15-20 cm absolute accuracy in the US, with claimed coverage work across Europe and Asia. The commercial packaging was the **Atlas DevKit** and **Atlas Lite DevKit** — car-mounted hardware units with a companion software suite intended to let developers create and crowdsource HD maps. Neither DevKit shipped with a public API or SDK on a public registry. Adjacent applications the company wrote about: road and pavement condition surveys, pipeline monitoring, insurance claims geospatial intelligence, smart rail and smart GIS. ## Writing - [Updates](https://civilmaps.com/category/updates/): company and engineering blog (posts run 2015-2019) - [White papers](https://civilmaps.com/category/white-papers/) - [Medium: @CivilMaps](https://medium.com/@CivilMaps): includes "To Build an Atlas" by CEO Sravan Puttagunta - Representative engineering posts: [a scalable approach to feature extraction from point cloud data](https://civilmaps.com/a-scalable-approach-to-feature-extraction-from-point-cloud-data/), [cognition for autonomous cars using 6D localization](https://civilmaps.com/cognition-for-autonomous-cars-using-6d-localization/), [moving to ARM: a marriage of minimums](https://civilmaps.com/moving-to-arm-a-marriage-of-minimums/), [on the accuracy of manual vs automated point cloud annotation methods](https://civilmaps.com/on-the-accuracy-of-manual-vs-automated-point-cloud-annotation-methods/) ## Funding and acquisition - Roughly USD 17 million raised across four rounds. - USD 6.6M seed led by Motus Ventures with Ford Motor Company, Wicklow Capital, StartX (Stanford) and AME Cloud Ventures — [announcement](https://civilmaps.com/news/civil-maps-closes-seed-funding-round-led-by-ford-motus-ventures/). - Acquired by [Luminar Technologies](https://www.luminartech.com/) in Q2 2022; announced at CES 2023-01-04. ## Code - [GitHub: civilmaps](https://github.com/civilmaps): organization created 2015-10-26, five public repositories, all dormant. - [cm-hal](https://github.com/civilmaps/cm-hal): the only first-party repo — a C++ hardware abstraction layer for LiDAR packet I/O, last pushed 2017-12-21. Source, not a distributable package. - `velodyne`, `loam_velodyne` and `libpointmatcher` are forks of third-party open-source robotics projects, not Civil Maps software. - No packages are published on npm, PyPI, RubyGems, crates.io, Packagist or NuGet. See `packages/civil-maps-packages.yml`. ## What this repository contains - `well-known/civil-maps-well-known.yml`: the full `/.well-known/` and contract-discovery probe record, including the soft-404 control test - `packages/civil-maps-packages.yml`: registry search record (no first-party packages) - `conformance/civil-maps-conformance.yml`: API standards recorded as not-applicable, plus corporate status - `security/civil-maps-domain-security.yml`: TLS/HSTS/DNSSEC/CAA/SPF/DMARC probe of civilmaps.com - No `a2a/` artifact: the agent-card probe returned an HTML soft-404 on both the canonical and legacy paths and was rejected.