Mapbox’s API Costs Are Quietly Pushing Devs Toward Felt

The Quiet Pricing Revolt in the Mapping World

Mapbox built its reputation as the developer-friendly alternative to Google Maps – flexible, customizable, and priced in a way that rewarded builders who were just getting started. That reputation has eroded. Since Mapbox restructured its pricing in 2021, moving away from a flat monthly model toward per-load and per-request billing, a growing segment of developers have found their bills climbing in ways that are difficult to predict and even harder to budget for. The platform’s power hasn’t diminished, but the cost of accessing it has changed the math for independent developers, small teams, and scrappy startups who once considered it a default choice.

Into that gap has stepped Felt, a collaborative mapping platform that launched in 2021 with a design-forward, browser-based approach to building and sharing maps. Felt doesn’t try to replicate everything Mapbox does – it doesn’t need to. It targets a specific frustration: that the process of making a useful, shareable map has become unnecessarily expensive and technically demanding. For the developers, data journalists, urban planners, and product teams who just need maps that work and don’t send finance into a spiral, Felt is starting to look very attractive.

How Mapbox’s Billing Model Became the Problem

Mapbox charges based on map loads and API requests, which sounds straightforward until a product scales or traffic spikes unexpectedly. A developer building a consumer-facing app with moderate traffic can move from a manageable monthly bill to something significantly larger without any change to their product – just more users doing what users do. That unpredictability is the core complaint. Mapbox’s pricing page has never been considered simple reading, and the per-request structure means that cost is tied to engagement, not a fixed overhead a team can plan around. For bootstrapped founders or solo developers managing tight margins, that’s a structural problem.

The 2021 pricing overhaul was not subtle. Mapbox moved from a model that many developers found reasonable into tiered structures with higher per-unit costs that hit harder for apps using dynamic maps, geocoding, navigation, or real-time data layers. Feedback surfaced across developer forums, Reddit threads, and GitHub issues – not a coordinated backlash, but a sustained, ongoing irritation. Some teams reported rebuilding parts of their stack to reduce API calls. Others started evaluating alternatives for the first time. The conversation shifted from “which features does Mapbox have” to “what’s the cheapest way to keep using maps at all.”

What Felt Actually Offers

Felt’s pitch is not purely about cost – though pricing is friendlier for many use cases. The platform is built around collaboration, with a Google Docs-style interface that lets multiple people edit and annotate a map in real time. For teams that need to share spatial data without asking every stakeholder to learn GIS software or parse API documentation, that accessibility matters. Felt supports uploading CSVs, shapefiles, and GeoJSON, and renders them visually without requiring a developer to write any code. That alone covers a significant chunk of what most non-engineering users actually need maps for.

The platform also handles larger datasets more gracefully than many teams expect from a browser-based tool. Felt can ingest files with millions of rows and still render them at interactive speed, which puts it in a different category from basic tools like Google My Maps. For data journalists mapping election results or city planners visualizing rezoning proposals, that performance ceiling matters. It’s not a toy version of a real mapping tool – it’s a different kind of real mapping tool.

Where Felt intentionally does not compete is in programmatic, production-grade map embedding. If a developer needs a fully customized map tile layer baked into a native app with routing logic and real-time user location tracking, Felt is not the answer. Mapbox, despite the billing friction, still owns that space along with Google Maps Platform. Felt’s founders have been clear that they’re building for a different user – one who prioritizes speed and collaboration over pixel-level cartographic control.

Felt’s free tier is generous by current standards: teams can create and share maps without hitting a paywall immediately, which mirrors how developer tools have learned to grow – let people experience value before asking for a credit card. That strategy has worked for tools like Figma, Notion, and Linear, and it fits the way Felt’s target users discover software. Someone shares a Felt map in a Slack channel, the recipient clicks through, makes an edit, and suddenly there’s a second user in the system.

The Developer Migration Is Still Early

The shift toward Felt is not a mass exodus – Mapbox still processes maps at enormous scale and remains deeply embedded in applications across logistics, real estate, and consumer products. What’s happening is more selective: developers who were on the margin of Mapbox’s value proposition, who never needed the full feature set but defaulted to it because it was the known quantity, are now reconsidering. The friction of an unexpected invoice is often the moment a team finally evaluates alternatives they had been putting off.

This pattern shows up repeatedly in the software tool market. A dominant platform raises prices or changes its model, and the developers who were always slightly underserved start looking around at the same time. That synchronized attention is exactly how challengers get distribution. It doesn’t require Mapbox to become unusable – it just requires the cost-benefit calculation to shift enough that switching feels worth the effort.

Where This Leaves the Mapping Stack

Felt is not the only alternative getting attention. OpenStreetMap-based tools like Leaflet and MapLibre – an open-source fork of Mapbox GL JS – have seen increased adoption from developers who want to escape proprietary billing entirely. MapLibre in particular gained traction after Mapbox changed the license on its GL JS library in 2020, prompting a community fork that now has active corporate contributors. For developers comfortable with self-hosting and configuration, the open-source route sidesteps pricing anxiety altogether.

Felt sits in a different position than these open-source options. It’s a managed product with a clean interface, not a library you wire up yourself. That distinction matters for the non-developer stakeholder who needs to be part of the mapping workflow. A city government communications team, a nonprofit research division, or a newsroom graphics desk can use Felt without a developer in the room. That accessibility is a real competitive advantage against tools that require engineering time just to set up.

The question hanging over Felt’s growth is whether its collaborative, design-focused model can expand into enough use cases to build a durable business – or whether it remains a niche tool for a specific kind of user while Mapbox continues dominating production applications. Felt raised a $25 million Series A in 2022, which gives it runway to develop, but mapping infrastructure is not a cheap category to compete in. The more interesting near-term pressure point is whether Mapbox adjusts its pricing again in response to visible developer frustration, or whether it calculates that enterprise and growth-stage customers make the economics work regardless of what indie developers think.

For now, every time a developer posts a Mapbox invoice screenshot in a Discord server and someone replies “have you tried Felt,” the conversion funnel gets a little shorter.