Discover the Best AI Tools

TypeScript library and read-only MCP server for financial day counts, accrued interest, business days and payment schedules. Calendars use coded rules and pinned exceptional closures that require updates.

Native terminal multiplexer keeps sessions in a background server and optionally ranks palette commands with Jev. Local search works without a key; semantic ranking requires provider access, and Windows uses a WSL2 server.

Chrome side panel uses Jev to select and fill tools exposed by a page’s WebMCP schemas. Requires WebMCP support and a TypeSafe key; state-changing calls need confirmation, and array arguments currently populate only their first element.

Argentum provides MCP tools for agent reputation, attestations and signed action trails anchored on-chain. Trails omit original action inputs; reproducing an action therefore depends on the operator retaining them. Chain receipts document recorded claims and integrity, not automatic proof of correct behavior or regulatory compliance, and chain operations require infrastructure and fees.

CashPilot MCP exposes earnings, worker status and logs from an existing CashPilot instance, plus service deployment and lifecycle controls. It needs that instance's admin API key and supports HTTP or stdio. Remote exposure requires deliberate authentication configuration; this bridge neither creates passive income nor guarantees returns.

M-Pesa MCP connects agents to Safaricom Daraja payment prompts and status queries, plus Africa's Talking SMS and airtime. It requires provider credentials and callback configuration, with sandbox and production modes. Production actions can transfer value or send messages; documented validation and logging are not independent security certification.

Indexes local folders for lexical, vector and hybrid search through MCP. Default vectors use lexical feature hashing rather than neural embeddings; neural models and Jev reranking are optional additions.

Connects coding agents to Dodo payment operations and searchable documentation through remote MCP servers. Payment actions require OAuth and can change real accounts.

Python automation layer separates code-owned observation, execution and verification from a model’s single bounded action choice. Requires configured browser or desktop adapters and provider access; caller-written checks determine whether a task really succeeded.

Minimal runnable reference skill lets Jev select bounded browser operations from observed controls. Requires browser setup and model credentials; it targets ordinary pages and omits the fuller inspection and guard features of the upstream implementation.

CLI and MCP browser layer lets Jev choose Chrome actions while a calling agent supplies goals and literal text. Requires a compatible model provider and browser; benchmarks cover limited tasks and uncertain runs return control for review.

Typed SDK, persistent CLI and MCP combine native Playwright actions with Jev semantic selection and assertions. Native operations need no model key; semantic features require a decision backend, and inconclusive model evidence should not be treated as success.

Jev Mobile persists Android tasks in SQLite and runs a single worker that observes, decides, journals, mutates and verifies a USB-connected device. It needs Android/ADB access and TypeSafe credentials. Validated scope focuses on notes, checklists and recovery; sensitive actions and unknown permission dialogs escalate rather than run autonomously.

Computer Use Jev is a Go/Swift macOS driver with optional Jev decision loops over observed accessibility targets. Direct scripted actions need no Jev; goal execution needs TypeSafe and optional planner credentials. It requires macOS permissions and build tools; low-confidence actions stop, but valid typed choices do not prove semantic correctness.

CHP MCP exposes spending gates, argument-bound approval receipts and a structured authorization/deny ledger. Its finance examples are synthetic orders, transfers and rebalances, not live bank or exchange connections. Production hosts must bind identity, secrets and human approval themselves; installing an allowed tool does not authorize its use.

A Python framework ships runnable decision-model agents for browser, desktop, simulated robotics and games using Jev or supported local models; remote providers need credentials and local models need their runtime dependencies, with benchmark gains limited to the reported tasks.

Library, CLI and MCP server use Jev to choose Playwright actions while a calling LLM supplies goals and text. Requires TypeSafe credentials and Chromium; multi-step goals should be split and counting or sorting outcomes independently checked.

CLI and daemon let agents control an isolated Playwright Chromium browser through Jev decisions. Requires Node, browser installation and provider credentials; navigation allowlists are not a network firewall and sensitive actions still need supervision.

Caches model responses locally through a Node cache or OpenAI-compatible Python proxy, with optional Jev semantic reuse. Requires upstream credentials for misses; replaying a response does not repeat an action or verify that the external state is unchanged.

Jev Desktop adds a scoped Jev decision loop to an existing Codex Computer Use runtime for browser tabs and native macOS apps. It needs that runtime and TypeSafe access; prepared field values stay local while selected labels and goals leave the machine. Consequential actions return to Codex, inaccessible widgets remain limited, and end-to-end speedup is unproven.

Arc CUA combines a standalone macOS accessibility driver with an optional decision-model loop for bounded desktop/browser subtasks. The driver needs no model key; the loop needs a provider, and native desktop backends are macOS-only. Accessibility/OCR coverage and model-based completion judgments can fail; this is not unrestricted autonomous control.

A CLI and Python SDK provide persistent virtual computers for agent code, browser and desktop work, locally or in Celesto Cloud; local use needs virtualization, cloud use needs credentials, and Windows/macOS guests have host-specific and preview limitations.

A Rust CLI searches code and text using local BM25 candidates followed by Jev relevance judgments, returning ranked files and line ranges without a vector database; it requires a paid TypeSafe API key and retrieval remains limited by candidate recall.

Hooks for Claude Code and Codex use configurable rules and Jev verification to warn or block tool calls and stopping decisions. They require service credentials and fail open when verification fails, including authentication problems; agent-editable rules and model judgments are not a hard security boundary.