Overview
Glitch made a running web app the starting point for programming. Its browser editor, automatic deployment and remixing let people change real code before learning server administration. That convenience attracted a creative community and an acquirer. It also left Glitch responsible for running and defending the resulting applications. By 2025, management said costs had grown while newer platforms reduced its distinctiveness. Hosting ended that July. [1] [2] [3]
The useful tension is between removing work for creators and accepting that work as a service obligation. Glitch succeeded at the first. Public evidence cannot establish whether a different price, architecture or ownership arrangement would have financed the second.
Image 1 / 1
Founding Story
The company began as Fog Creek Software in New York in September 2000, founded by Joel Spolsky and Michael Pryor. Spolsky previously worked on Microsoft Excel’s Visual Basic for Applications and at Juno. The reviewed primary material does not establish Pryor’s earlier credentials or how the two founders first met. Their company later developed FogBugz and Trello. Spolsky’s 2014 account explicitly separates Trello Inc. from Fog Creek; their later outcomes must remain separate. [4] [5]
Glitch arose much later. Designer Pirijan’s account dates the original proposal with DanielX to 2015 and describes an internal product experiment. HyperDev entered public beta in May 2016, became Gomix, and then Glitch. Fog Creek adopted the Glitch company name in September 2018. These are product and company milestones, rather than several new company foundings. [6] [7]
Spolsky wrote at launch, “We want to eliminate 100% of the complicated administrative details”. In December 2016, when Anil Dash became CEO, Spolsky explained another need: “we need his soapbox and industry connections”. The first quote describes the product’s promise; the second shows why developer-community leadership mattered alongside engineering. Dash had worked with Six Apart and founded ThinkUp, Expert Labs and MakerBase. He was an experienced outside CEO, rather than an original Fog Creek founder. [1] [8]
Timeline
- September 2000: Spolsky and Pryor founded Fog Creek Software. [4]
- May 2016: HyperDev entered public beta. [1]
- September 2018: Fog Creek became Glitch Inc. [7]
- December 2018: Glitch announced a migration partnership with Mozilla’s Thimble. [9]
- July 2020: Glitch explained its free project-hour limits and paid Boosted Apps. [10]
- March 2021: CWA announced the workers’ first collective bargaining agreement. [11]
- May 18, 2022: Fastly acquired all voting equity in Glitch. [12]
- June 2024: Glitch previewed sharing externally hosted projects on paid members’ profiles. [13]
- July 2025: Hosting ended; a July 23 team post said most wind-down work was complete. [2] [3]
What They Built
A visitor to HyperDev received an already running site, a URL and editable client and server files. Typing changes triggered deployment and browser refresh. The server ran Node.js, so users could make functioning applications rather than only static visual demonstrations. Collaboration and remixing made another person’s project a starting point. [1]
The team learned where convenience became confusion. Pirijan described dropping an initial language/framework selection screen because it delayed creation. A code-snippet feature confused testers and raised maintenance questions across files and packages. It became a more predictable Add Packages button. Hiding initialization code also made some users less comfortable. These decisions show an effort to remove setup while keeping the resulting application understandable. [6]
Glitch also operated a discovery and teaching community. Mozilla’s Thimble transition offered a practical distribution channel: Glitch committed to maintaining more than 1.4 million Thimble projects and planned migration around academic semesters. This was a commitment and project inventory, rather than a verified count of active or paying users. [9]
Market Position
Target Customers
The product served learners, teachers, creative coders and developers making small applications. Company case material describes artist Naoto Hiéda paying to keep a virtual exhibition available continuously and gain storage. That is a specific reason to upgrade: an experimental project became part of a public portfolio. It does not establish how often the wider free community converted. [14]
Market Size
Fastly reported more than 1.8 million developers at acquisition. The May 2025 announcement described millions of users and tens of millions of apps. These cumulative company figures measure different things; neither establishes current active creators, paying accounts or an obtainable revenue market. [15] [2]
Competition
Glitch initially compressed editing, deployment and sharing into one interaction. Competitors later offered overlapping convenience. Val Town’s May 2025 migration guidance explicitly credited Glitch, offered browser editing and instant reload, and warned that its TypeScript environment was not a drop-in replacement for every Glitch app. Its current pricing still includes a free tier. A new entrant must compare the actual workload and migration requirements, rather than assume browser hosting is missing. [16] [17]
Business Model
Glitch combined free access with paid capacity. In July 2020, static sites no longer slept; free dynamic projects had a project-hour allowance. Boosted Apps offered up to five projects with continuous operation and more memory, storage and higher limits. The commercial boundary followed resource consumption and reliability needs. [10]
TechCrunch reported a $30 million Tiger Global Series A that closed in November 2018. [18] Fastly’s filing records a separate $34.9 million cash acquisition in May 2022, including an $8 million indemnification holdback. Acquisition consideration is not Glitch revenue or a measure of investor return. Fastly did not present separate pro forma operating results because Glitch was immaterial to the consolidated accounts. Standalone margins, revenue and paid conversion remain unavailable. [12]
Traction
A community can generate value beyond subscriptions. Fastly’s acquisition announcement placed the Glitch team in a new Developer Experience organization producing templates, tools, tutorials and feedback channels. That is a concrete buyer rationale. It leaves the eventual effect on Fastly product adoption unmeasured. [15]
Community operations continued after acquisition. The June 2024 roundup documented a Rainbow Code Jam and invited projects under #MadeWithGlitch. Workers also secured formal protections: CWA’s March 2021 announcement describes grievance, severance and recall provisions. The reviewed evidence does not connect the union to the hosting decision. [19] [11]
Outcome and evolution
By October 23, 2026, the original app-hosting service had ended. The May announcement set July 8, 2025 as the cutoff; the July 23 team post confirmed that ending hosting support was mostly complete. The root domain now redirects to the surviving blog. The earlier promise of redirects through at least the end of 2026 does not verify every project redirect today. [2] [3] [20]
Management cited operating costs, abuse and reduced distinctiveness. A competing explanation is that an acquirer deprioritized a small service. Fastly had originally promised continued collaboration and technical integration; Dash later explicitly supported the hosting decision. The observed record establishes the decision and stated reasons, but cannot isolate internal budget allocation from the service’s own economics. [2] [15]
The 2024 external-project preview had already separated discovery from hosting. That experiment matters: community value did not logically require operating every application. The later retirement of profiles means it did not preserve the original community product either. [13] [2]
Key Lessons
- Glitch’s early removal of language choices and confusing hidden code made creation easier. Test which abstractions reduce work without making behavior harder to understand.
- Thimble migration brought projects and an educational mission, together with continuing hosting obligations. Measure the resources and revenue associated with imported activity.
- Fastly bought developer access and tools; that did not guarantee indefinite hosting. Product continuity needs its own operating commitment after an acquisition.
- Glitch offered exports and redirects during retirement. Source portability and URL continuity require separate plans because preserving code alone does not preserve a running app.
Sources
- Spolsky, Introducing HyperDev, May 2016
- Glitch hosting announcement, May 2025
- Glitch completed wind-down, July 2025
- Spolsky official biography
- Spolsky on the separate Trello company, July 2014
- Glitch product design history, March 2017
- Fog Creek company rename, September 2018
- Spolsky appoints Anil Dash, December 2016
- Glitch and Mozilla Thimble migration, December 2018
- Glitch free hours and Boosted Apps, July 2020
- CWA agreement announcement, March 2021
- Fastly 2022 annual filing, business combinations
- Glitch external-project preview, June 2024
- Glitch artist customer case, April 2020
- Fastly acquisition announcement, May 2022
- Val Town migration guide, May 2025
- Val Town current pricing, checked October 2026
- TechCrunch financing report, July 2019
- Glitch community roundup and campaign, July 2024
- Glitch root-domain redirect, checked October 2026