Open Source Coding CLI for teams that need more control
Built for developers who want strong coding performance without being locked into a single model, provider, or environment.
Coding agents are becoming infrastructure
Most coding tools are built around a single hosted model, a closed workflow, and limited control over how context, memory, tools, and execution are handled.
That may be enough for individual experimentation
It becomes limiting when teams need reproducibility, model flexibility, private deployment, secure environments, or control over cost and performance.
R-CLI gives teams an open coding harness they can inspect, run, extend, and deploy where they need it.
#1 on Terminal-Bench 2.1
R-CLI solved 75 of 89 tasks โ 84.3% โ running Claude Opus 4.8. That is the first result past 84%, and 5.4 points clear of the next-best harness on the same model: the gain comes from the harness, not the model.
Terminal-Bench 2.1 โ 89 tasks, Daytona environment, run 3 July 2026. Figures for other agents are from the public Terminal-Bench 2.1 leaderboard. Full logs, methodology and per-task verifier reports are published.
Built for high-performance AI software development
01.
Model flexibility
Run the models that fit your requirements across frontier, open-source, local, or private deployments.
Choose based on quality, latency, privacy, cost, or hardware availability.
02.
Context and memory
Give coding workflows persistent context across tasks, repositories, sessions, and iterative development cycles.
Reduce the need to repeatedly re-explain codebases, architecture, constraints, and prior decisions.
03.
Tool use and execution
Enable agents to inspect files, modify code, run commands, test changes, and work through multi-step software tasks.
04.
Open and extensible
Inspect the harness, adapt workflows, add tools, change model providers, and integrate R-CLI into your own development environment.
05.
Private and sovereign deployment
Run coding workflows in cloud, private cloud, on-premise, or air-gapped environments depending on your organizationโs requirements.
Start free, scale with usage.
$0
USD / month
Free
โข Backboard API access
โข R-CLI access
โข Web/mobile workspace
โข Basic memory and RAG access
โข Basic usage analytics
โข No credit card required
POWERED BY
$5
USD / month
Education
โข Everything in Free
โข Lower-cost access for students and researchers
โข Memory and RAG access
โข Active Context Management
โข Usage analytics
โข Tokens billed at provider price
$20
USD / month
Developer
โข Everything in Education
โข API, R-CLI, desktop, and web/mobile access
โข Advanced memory and context tools
โข Workflow and usage analytics
โข BYOK support
โข Tokens billed at provider price
$400
USD / month
Scale
โข Everything in Developer
โข Team workspace access
โข Higher usage capacity
โข Advanced analytics
โข Shared billing and usage visibility
โข Priority support
โข Production workflow support
Proven in real coding environments
R-CLI has demonstrated leading performance on Terminal Bench, with reproducible logs, methodology, and system details published publicly.
The goal is not simply to generate code snippets. It is to help AI systems operate inside real development environments where they must inspect codebases, use tools, execute commands, test changes, and complete multi-step tasks.
Open source, with a path to enterprise scale
R-CLI is designed to stand on its own as an open-source tool.
Teams can use it locally, inspect the code, adapt it to their workflows, and run it against the models and infrastructure they choose.
When they need persistent memory, model routing, deployment controls, team-level governance, or private infrastructure, they can connect R-CLI to the Backboard Platform.
This gives developers an open starting point and organizations a clear path to production-grade infrastructure.
Built for teams that cannot compromise
R-CLI is designed for:
01
Engineering teams building production software with AI
02
Organizations with private or regulated codebases
03
Teams evaluating multiple coding models
04
Companies reducing dependency on hosted coding tools
05
Developers working in restricted or disconnected environments
06
Enterprises that need auditability, control, and deployment flexibility
Build software with AI on
infrastructure you control.
Use an open coding harness today. Add sovereign AI infrastructure as your requirements grow.