Stencil Labs, Inc
Stencil Labs builds new tools to make agentic software development more seamless. It describes its goal as creating a collaborative ‘whiteboard’ where engineers, reviewers, designers and managers can work with generative models without sacrificing quality or individual judgment.
At a Glance
- Software engineers and developers
- Teams adopting coding agents
- Organizations building agentic development workflows
- Researchers and open-source developers working on AI harnesses
- +1 more
AI Tools by Stencil Labs, Inc
(2)ovnfs
Copy on Write Folders for Agents
omp (oh-my-pi)
Terminal AI Coding Agent CLI
Discussions
No discussions yet
Be the first to start a discussion about Stencil Labs, Inc
Latest News
The Harness Playbook: Stencil published an architectural postmortem and playbook for building durable, multi-client agentic coding harnesses.
Prewalk: research showing a frontier-model trajectory can let a cheaper model retain 92–97% of frontier coding performance at about half the cost.
Snapcompact: a local image-based context-compaction technique reporting near-verbatim recall at roughly one-third of input price.
Hashline research: content-hash line anchors improved coding success over patch for 14 of 16 tested models.
Products & Services
Open-source coding-agent harness with an IDE wired in, powered by a native Rust core and tools tuned against real sessions. It supports interactive, one-shot, RPC and ACP entry points, and can be embedded through an SDK.
Open-source infrastructure for agents: isolated, copy-on-write views of any folder or Git repository that can be reviewed with diff and merged with apply. Checkouts can be snapshotted, forked and attached to another machine with lazy content transfer; it serves mounts through NFSv3 on macOS, FUSE on Linux, and SMB or WinFsp on Windows. Licensed under MIT.
An edit format that tags file lines with short content hashes, allowing models to reference stable line anchors instead of reproducing old text and rejecting edits when the file has changed.
A research technique in which a frontier model explores a task and provides a trajectory for a cheaper model to execute, aimed at reducing reading cost while preserving coding performance.
Market Position
Stencil positions itself as a research-led, model-agnostic alternative to vendor-specific coding-agent harnesses. Its differentiation is in improving the harness layer—state, tools, editing formats, context handling and workspace isolation—rather than competing primarily by training a model; Hashline is explicitly compared with patch, string replacement and Cursor-style edit systems in its research.
Leadership
Founders
Can Bölük
Founder and CEO; systems engineer and security researcher with a decade of experience across kernels, reverse engineering and compilers. Leads product direction and R&D.
Meehir Patel
Co-founder and COO; engineering leader and technical director with experience running engineering organizations. Leads partnerships and enterprise adoption from San Francisco.
Executive Team
Can Bölük
Founder & CEO
Systems engineer and security researcher with a decade across kernels, reverse engineering and compilers; leads product direction and R&D.
Meehir Patel
Co-founder & COO
Engineering leader and technical director experienced in running engineering organizations; leads partnerships and enterprise adoption from San Francisco.
Founding Story
Stencil says it was started to overcome the friction that still makes it difficult to get reliable results from generative models, even for domain experts. Its initial vision is to build tools and a research-oriented environment that lets people collaborate with agents while preserving software quality and human touch.
Business Model
Revenue Model
No commercial revenue model is stated on the reviewed official pages. The company presents itself as a research company publishing open-source research and tools; omp is available as an installable coding agent and ovnfs is MIT-licensed open source.
Target Markets
- Software engineers and developers
- Teams adopting coding agents
- Organizations building agentic development workflows
- Researchers and open-source developers working on AI harnesses
- Enterprise engineering organizations evaluating agent automation
- Interactive coding agents and IDE-assisted software development
- Autonomous software factories and untrusted repository workflows
- Multiple agents or subagents working in a shared workspace
- Remote driving or observation of coding agents
- Reviewing and merging agent changes without modifying the source workspace
- Reducing inference and context costs on long-horizon coding tasks