Grove manages project scripts through a Mac menu-bar app, CLI and built-in MCP server controlling the same terminal processes. It discovers supported manifests, shows live logs and offers on-device error summaries.
GitWarren reviews staged, unstaged and untracked changes directly from a Git worktree before a pull request. It offers a desktop or browser UI and attributed agent participation through a local MCP server.
Clockwork schedules existing coding-agent CLIs on a Mac or approved paired worker, using worktrees, budgets and run reports. It supports calendar recurrence, event triggers and review before opening a pull request.
SODAX provides SDK and API interfaces for cross-network asset quoting and execution. Its Builders MCP helps coding agents navigate integration, while simplified APIs support narrower execution workflows.
Replay QA explores application journeys, records runtime sessions and reports bugs with investigation context and suggested fixes. Its security scans cover issues such as injection, access control and business logic.
HyperProbe integrates with coding assistants to capture read-only variable state at selected lines in a running service. It connects incident evidence with investigation, using configurable probe expiry and rate limits.
dif.sh stores feature flags and experiments as reviewable Markdown files in a repository. Its CLI builds runtime configuration and agent context, while optional Dif Cloud collects metrics and reports outcomes.
CosmoDex combines programming courses, practice and coding challenges with gamified progression. Its Pro scope advertises AI code review and analytics, while several courses and competition features remain forthcoming.
Coldtea connects coding-agent sessions with tasks, pull requests, agentic QA and production monitoring. Its desktop workspace includes terminals, worktrees and review context, with optional cloud runs.
Airuncode runs coding agents on a desktop machine with user-supplied model accounts, API keys or local models. It includes parallel agents, test workflows, context snapshots and an optional linked-device PWA.
Ruby UTCP implements Universal Tool Calling Protocol 1.1 for discovering, storing and calling tools directly over their native transports. It provides structured call templates, authentication and protocol-specific errors.
TryCase runs an application against the user journeys affected by a pull request and reports a verdict with video and screenshots on GitHub. It sizes runs by changed source files and quotes the charge before testing.
QApilot MCP automates local Android devices and emulators from an MCP-compatible AI client. It records steps, previews the app, saves test cases and produces test reports using an Appium-based setup.
Modeinspect connects a design canvas with a codebase and agents for live capture and design review. It helps teams inspect interface changes and prepare reviewable implementation work.
Juggler is a coding-agent interface with a native Go server, synchronized conversations and inspectable tool calls. It supports local and remote clients, direct model-provider connections and an extension SDK.
Jackalope is an open-source desktop coding workspace for parallel agents. It separates work into Git worktrees and supports browser tools, MCP connections and review of proposed changes.
hob is a desktop workspace for working with existing AI agents, files, browsers, terminals and reviewable changes. It supports parallel work, worktrees and automations, with remote access on eligible plans.
Devin Voice lets users discuss coding work aloud with Devin while navigating the application. The documented controls distinguish muting the microphone from silencing the agent’s audio, and conversation history remains available.
SWE-2 is Cognition’s coding model for multi-step software engineering tasks. The official release describes multiple reasoning-effort levels and availability through Devin Desktop and CLI, with rollout to Devin Web and Fusion.
chat-recall stores coding conversations for full-text and knowledge-graph retrieval. It offers MCP access, analytics and code-related findings, with hosted and self-hosted deployment options.
Harden Agentic Integrity Foundation checks supported coding-agent tool calls before execution. It offers local decision history, blocking and safe-retry workflows, using native hooks or an MCP proxy fallback.
Yedric adds a tool-calling assistant inside an application, using its documentation, APIs and configured actions. It also records user questions and actions for product analysis.
Twigg stores conversations, assembles context for each model and routes requests across supported LLM providers. Applications create a chat and send subsequent events rather than resending full conversation history.
tiun combines authentication, checkout, customer records and analytics for AI and SaaS products. It provides merchant-of-record payments with SDK and MCP integration and support for subscriptions and usage-based pricing.
2.9% + $0.30/transaction
API unknown
Checked Oct 9, 2026
Showing 240 of 2,280 tools
Plan facts carry individual sources and check dates. Unknown values are excluded from required filters.