Playgama launches AI agent publishing and improvement tools for browser games

Playgama launched Playgama Sandbox, MCP and an updated Bridge SDK for AI agents. The tools let agents publish browser games and analyze player data automatically. Agents then ship updated builds without a human live-ops team. This closes the feedback loop between real player data and iterative game design.

Playgama MCP uploads a game and returns a live, shareable link in minutes.
The updated Bridge SDK adds event-level analytics on player behavior.
Agents pull this data through MCP to find where players drop off.
They then push a new build without any developer action.

Titles that retain players move to a network of 450 million+ monthly players.
That reach spans 15+ instant-play platforms, including YouTube Playables.
Developers earn ad revenue from launch and a path to paid user acquisition.
"The next opportunity is helping agents make better decisions about the second, third, and tenth versions." - Dmitry Kachmar, Founder, Playgama

HTML5 games run entirely in code, easier for agents than binary engines.
GPT-6 Astra and Claude Opus 5.5 fueled a surge of AI-made games in September.
Playgama positions Sandbox as the discovery layer before titles reach its main catalog.

For Publishers & Developers: Greenlight HTML5 concepts only after agent-iterated builds prove retention with live players.
For Investors & VC: Add agent iteration speed and per-build retention lift to web-game studio diligence.
For Service Vendors: Redirect live-ops and QA outreach toward engine-based clients, where agent automation still lags.