npm - apteva - Versions diffs - 0.30.0 → 0.30.2 - Mend

apteva 0.30.0 → 0.30.2

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Files changed (2) hide show

package/README.md +94 -182
package/package.json +1 -1

package/README.md CHANGED Viewed

@@ -1,244 +1,156 @@
 <p align="center">
-  <img src="https://apteva.ai/icon.png" width="60" />
+  <a href="https://apteva.ai">
+    <img src="https://apteva.ai/icon.png" width="96" alt="Apteva logo" />
+  </a>
 </p>
 <h1 align="center">Apteva</h1>
 <p align="center">
-  <strong>An AI that never stops thinking.</strong>
+  <strong>AI agents, batteries included.</strong>
 </p>
 <p align="center">
-  Apteva is a continuous thinking engine. It doesn't wait for prompts —<br>
-  it observes, reasons, decides, and acts on its own initiative. Around the clock.
+  The open-source, self-hosted platform for agents that keep working after the chat ends.
+</p>
+<p align="center">
+  Give agents goals. They remember context, react to events, schedule and resume work,<br />
+  delegate to workers, and operate through apps and 575+ integrations.
 </p>
 <p align="center">
   <a href="https://apteva.ai">Website</a> ·
-  <a href="https://apteva.ai/features">Features</a> ·
+  <a href="https://docs.apteva.ai/get-started">Docs</a> ·
+  <a href="https://apteva.ai/apps">Apps</a> ·
   <a href="https://apteva.ai/cloud">Cloud</a> ·
-  <a href="https://github.com/apteva/core">Core Engine</a> ·
-  <a href="https://github.com/apteva/server">Server</a> ·
-  <a href="https://github.com/apteva/computer">Computer</a> ·
-  <a href="https://github.com/apteva/integrations">Integrations</a>
+  <a href="https://discord.gg/apteva">Discord</a>
+</p>
+<p align="center">
+  <a href="https://github.com/apteva/apteva/releases/latest"><img src="https://img.shields.io/github/v/release/apteva/apteva?style=flat-square" alt="Latest release" /></a>
+  <a href="https://www.npmjs.com/package/apteva"><img src="https://img.shields.io/npm/v/apteva?style=flat-square" alt="npm version" /></a>
+  <a href="https://github.com/apteva/apteva/actions/workflows/release.yml"><img src="https://img.shields.io/github/actions/workflow/status/apteva/apteva/release.yml?style=flat-square&label=release" alt="Release status" /></a>
+  <a href="https://github.com/apteva/apteva"><img src="https://img.shields.io/github/stars/apteva/apteva?style=flat-square" alt="GitHub stars" /></a>
 </p>
 ---
-## Quick Start
+## Start in 30 seconds
 ```bash
 npx apteva
 ```
-A setup wizard guides you through provider selection, API key, and capabilities. Then the agent starts thinking.
+On macOS or Linux with Node.js 18+. The installer downloads the current native release. Connect a model provider, create an agent, and Apteva opens the local dashboard.
-Or from source:
+Apteva runs locally by default. Your agents, credentials, memory, and operational data stay under your control.
-```bash
-cd core && go build -o apteva-core .
-cd ../server && go build -o apteva-server .
-cd ../apteva && go build -o apteva .
-./apteva
-```
+## Agents are easy to demo. Hard to operate.
-Or with Docker:
+A prototype agent can call a tool. A production agent also needs durable state, memory, permissions, events, schedules, workers, retries, dashboards, integrations, and deployment.
-```bash
-docker run -p 5280:5280 -v apteva-data:/data apteva
-```
+Apteva packages that operating layer into one workspace. Instead of rebuilding the same infrastructure for every agent, you get a complete system for ongoing operations.
-## What is Apteva?
+| | What Apteva provides |
+|---|---|
+| **Keeps working** | Agents react to events, follow up on unresolved work, and continue across hours or days. |
+| **Durable by default** | Tasks, schedules, history, memory, and progress survive restarts and closed chats. |
+| **A complete operating layer** | Runtime, server, dashboard, channels, apps, integrations, files, logs, and deployment work together. |
+| **Apps, not just tool wrappers** | One app can add tools, UI panels, routes, workers, channels, memory, and domain workflows. |
+| **Multi-agent operations** | Agents delegate bounded work to dedicated workers while retaining durable ownership and visibility. |
+| **Control where it runs** | Use Apteva locally, self-host it on a VPS or Docker, or run it on Apteva Cloud. |
-Every AI you use is reactive — you type, it responds, then it stops. Apteva is different. It runs a **continuous thinking loop**: observe → reason → act → sleep → repeat. It wakes when events arrive, sleeps when nothing happens, and thinks proactively in between.
+## What agents can operate
-It's not a chatbot. It's not an agent framework. It's **artificial consciousness** — a machine that is always aware, always ready, whether anyone is watching or not.
+| Operation | Examples |
+|---|---|
+| **Customer support** | Triage tickets, search knowledge bases, draft replies, and escalate when human judgment is needed. |
+| **Sales and CRM** | Enrich leads, schedule follow-ups, update pipelines, and surface deals that need attention. |
+| **Content and growth** | Research, create assets, publish across channels, and monitor performance. |
+| **Engineering and DevOps** | Watch deployments, investigate alerts, run tests, prepare patches, and coordinate incidents. |
+| **Back-office work** | Process invoices, reconcile data, manage inventory, and coordinate vendors. |
+| **Devices and edge systems** | Run the Go-based agent core close to browsers, machines, robots, and local infrastructure. |
-### What it does between prompts
+## Apps turn agents into operators
-- Notices a pattern in your support tickets before you ask
-- Follows up on a cold lead because three days passed and it remembers
-- Reorganizes inventory because it spotted a trend in last week's orders
-- Spawns a worker thread to handle a customer email at 3am
-- Evolves its own directives as it discovers better approaches
-- Browses the web, takes screenshots, clicks through pages
-- Connects to GitHub, Stripe, Slack — 263+ integrations
+Apteva apps extend the platform with any combination of:
-## Architecture
+- MCP tools and integrations
+- Dashboard and chat UI
+- HTTP routes and webhooks
+- Background workers and scheduled jobs
+- Channels, memory, and shared operational data
-```
-┌──────────────────────────────────────────────────────┐
-│  ./apteva                                            │
-│  Setup wizard → TUI terminal interface               │
-│  Channels: CLI, Telegram, Discord                    │
-└──────────────────┬───────────────────────────────────┘
-                   │
-┌──────────────────▼───────────────────────────────────┐
-│  apteva-server (:5280)                               │
-│  Auth · Instances · Integrations · Webhooks          │
-│                                                      │
-│  ┌─────────┐  ┌─────────┐  ┌─────────┐              │
-│  │ core-1  │  │ core-2  │  │ core-3  │  ...         │
-│  │ default │  │ support │  │ monitor │              │
-│  └────┬────┘  └────┬────┘  └────┬────┘              │
-│       │            │            │                    │
-│       ▼            ▼            ▼                    │
-│  ┌──────────────────────────────────────────────┐    │
-│  │              MCP Tools                       │    │
-│  │  channels · exec · web · browser             │    │
-│  │  github · stripe · slack · 263+ more         │    │
-│  └──────────────────────────────────────────────┘    │
-│                                                      │
-│  memory · self-pacing · evolve · persistent history  │
-└──────────────────────────────────────────────────────┘
-```
+The current catalog contains **575+ integrations**, including GitHub, Slack, Stripe, Shopify, Airtable, Twilio, HubSpot, Google Workspace, cloud providers, databases, media tools, and model APIs.
-One command (`./apteva`) starts everything. The CLI spawns the server, the server spawns core instances, and the agent begins thinking.
-## Key Features
-| Feature | Description |
-|---------|-------------|
-| **Continuous Thinking** | Infinite loop — observe, reason, act, sleep. Not request-response. |
-| **Multi-Threaded** | Spawns worker threads for parallel tasks. Each has its own tools, pace, and directive. |
-| **Self-Evolving** | Persistent memory. Refines its own directives over time. Gets sharper the longer it runs. |
-| **Self-Pacing** | Sets its own sleep duration — 2 seconds when busy, hours when idle. Events wake it instantly. |
-| **Session Persistence** | Conversation history survives restarts. JSONL per thread, auto-compaction, never loses context. |
-| **Agent-Driven Safety** | No forced approval gates. The agent decides, learns from feedback, asks when unsure. Three modes: autonomous, cautious, learn. |
-| **263+ Integrations** | GitHub, Slack, Stripe, Shopify, and more. Each runs as its own MCP server. Credentials encrypted. |
-| **Browser Control** | Provided by the Computer app through MCP tools and app-managed sessions. |
-| **Multi-Channel** | CLI terminal, Telegram, Discord. Agent routes responses to the right channel. |
-| **Multi-Instance** | Run multiple agents in parallel. Each has its own directive, tools, and history. Projects for isolation. |
-| **Terminal UI** | Two-panel TUI with live status, thread thoughts, streaming responses, modal commands. |
-| **Embeddable** | The core is a standalone Go binary. Run it headless, connect via API, embed anywhere. |
-## Use Cases
-- **Run a business** — Support, sales, content, billing — fully autonomous, 24/7
-- **Content pipeline** — Research, write, publish, distribute, track — zero human input
-- **Ad operations** — Monitor ROAS, adjust budgets, generate creatives, A/B test
-- **DevOps** — Monitor deploys, triage alerts, write patches, coordinate incidents
-- **Personal assistant** — Email triage, calendar management, reminders, research
-- **Web automation** — Browse sites, fill forms, take screenshots, extract data
-## CLI Commands
-| Command | Description |
-|---------|-------------|
-| `/status` | Core status |
-| `/config` | Full config |
-| `/directive [text]` | Show or set directive |
-| `/mode` | Switch mode: autonomous, cautious, learn |
-| `/threads` | List/kill threads |
-| `/integrate <app>` | Connect an integration (263+ apps) |
-| `/connect telegram` | Connect Telegram bot |
-| `/channels` | List connected channels |
-| `/mcp` | Manage MCP servers |
-| `/help` | All commands |
-Everything else you type is sent to the agent.
-## Modes
-| Mode | Behavior |
-|------|----------|
-| **autonomous** | Agent acts freely. Learns from feedback. Trusted. |
-| **cautious** | Asks before destructive or external actions. Learns from answers. |
-| **learn** | Asks about every new tool type. Builds a safety profile over time. |
-All modes use `[[remember]]` for persistent learning. The agent gets smarter the longer it runs.
-## Providers
-| Provider | Models |
-|----------|--------|
-| **Fireworks** | Kimi K2.5 |
-| **Anthropic** | Claude Sonnet 4, Claude Haiku 4.5 |
-| **OpenAI** | GPT-4.1, GPT-4.1-mini |
-| **Google** | Gemini 2.5 Pro, Gemini 2.5 Flash |
-Select during setup. Switch anytime with `/mode` or the API.
+[Browse apps](https://apteva.ai/apps) · [Explore the integrations repository](https://github.com/apteva/integrations) · [Build with the App SDK](https://github.com/apteva/app-sdk)
-## Repositories
-| Repo | Description |
-|------|-------------|
-| [`apteva/apteva`](https://github.com/apteva/apteva) | This repo — CLI, setup wizard, TUI, npm launcher |
-| [`apteva/core`](https://github.com/apteva/core) | The thinking engine (Go) |
-| [`apteva/server`](https://github.com/apteva/server) | Management server — auth, instances, integrations, dashboard (Go) |
-| [`apteva/integrations`](https://github.com/apteva/integrations) | 263 app connectors + webhook registrar (TypeScript) |
-## Configuration
+## How it works
-```bash
-# Data stored in
-~/.apteva/apteva.json    # CLI config (capabilities, API key, instance ID)
-~/.apteva/apteva.db      # Server database (auth, connections, providers)
+```mermaid
+flowchart LR
+    U["Dashboard, CLI, and channels"] --> S["Apteva server"]
+    E["Events, webhooks, and schedules"] --> S
+    S --> A["Persistent agent runtimes"]
+    A --> W["Durable tasks and workers"]
+    S --> P["Apps"]
+    P --> C["Tools, UI, routes, channels, and memory"]
+    P --> I["575+ integrations"]
 ```
-## From Source
-```bash
-# Build all three binaries
-cd core && go build -o apteva-core .
-cd ../server && go build -o apteva-server .
-cd ../apteva && go build -o apteva .
+The CLI installs and starts the platform. The server manages authentication, projects, agent runtimes, apps, connections, events, and the dashboard. Each agent core owns its thinking loop, threads, tools, memory, and persistent history.
-# Run (starts server + core + TUI)
-cd apteva && ./apteva
+## Run it your way
-# Or run core standalone (headless)
-cd core && ./apteva-core --headless
+### Local
-# Or connect CLI to a remote server
-./apteva --no-spawn --server <host>:5280
+```bash
+npx apteva
 ```
-## Docker
+### Docker
 ```bash
-# Build
-docker build -t apteva .
+docker run -d \
+  --name apteva \
+  -p 5280:5280 \
+  -v apteva-data:/data \
+  ghcr.io/apteva/apteva:latest
+```
-# Run
-docker run -p 5280:5280 -v apteva-data:/data apteva
+Then open [http://localhost:5280](http://localhost:5280). Pin a numbered image tag instead of `latest` for production deployments.
-# Connect CLI to it
-./apteva --no-spawn --server <host>:5280
-```
+### Cloud
-## Migrating from OpenClaw
+[Apteva Cloud](https://apteva.ai/cloud) runs the same platform without managing the server yourself.
-Already running OpenClaw agents? Here's how to switch.
+## Bring your models
-| OpenClaw | Apteva | Notes |
-|----------|--------|-------|
-| `claw.run(prompt)` | `config.json` → directive | Apteva doesn't need prompts. It runs continuously. |
-| `claw.tool(...)` | MCP server | Any MCP server works. 263+ built-in. |
-| `claw.memory` | `[[remember]]` | Persistent across restarts. Embedding-based recall. |
-| `claw.agent(name)` | `[[spawn id="name"]]` | Threads are cheaper. They share memory and tools. |
-| Webhook handlers | Subscriptions | Events route directly to threads. |
-| `OPENCLAW_API_KEY` | `FIREWORKS_API_KEY` | Or Anthropic, OpenAI, Google. Your choice. |
-| Cron jobs | Self-pacing | No cron needed. `[[pace sleep="5m"]]` and it wakes on events. |
-| Agent state files | Evolving directives | The agent rewrites its own config. It improves itself. |
-| Session files | JSONL history | Persistent per thread, auto-compaction, survives restarts. |
+Use hosted or local models, choose different providers per project, and switch models without rebuilding your agents. Apteva supports OpenAI, Anthropic, Google, Fireworks, Ollama, NVIDIA, Venice, xAI, and other compatible providers.
-### Why switch?
+## Repositories
-- **OpenClaw agents sleep between requests.** Apteva agents think between requests.
-- **OpenClaw needs orchestration.** Apteva orchestrates itself.
-- **OpenClaw bills per call.** Apteva self-paces to minimize cost — sleeps long when idle, thinks fast when busy.
-- **OpenClaw agents forget.** Apteva agents remember, learn, and evolve their own directives.
+| Repository | Role |
+|---|---|
+| [`apteva/apteva`](https://github.com/apteva/apteva) | CLI, installer, local lifecycle, and releases |
+| [`apteva/core`](https://github.com/apteva/core) | Persistent agent runtime and thinking loop |
+| [`apteva/server`](https://github.com/apteva/server) | Management API, agent orchestration, apps, and embedded dashboard |
+| [`apteva/dashboard`](https://github.com/apteva/dashboard) | React administration and operations UI |
+| [`apteva/integrations`](https://github.com/apteva/integrations) | Integration catalog, OAuth, webhooks, and MCP generation |
+| [`apteva/apps`](https://github.com/apteva/apps) | First-party operational apps |
+| [`apteva/app-sdk`](https://github.com/apteva/app-sdk) | Go SDK for building Apteva apps |
+| [`apteva/computer`](https://github.com/apteva/computer) | Browser and computer-use backends |
-The fundamental difference: OpenClaw agents are **functions you call**. Apteva agents are **processes that run**. One waits. The other works.
+## Community
-## License
+Read the [documentation](https://docs.apteva.ai/get-started), join the [Discord community](https://discord.gg/apteva), or open an [issue](https://github.com/apteva/apteva/issues) for bugs and feature requests.
-[MIT](LICENSE)
+If Apteva helps you build agents that do real work, [star the repository](https://github.com/apteva/apteva) so more builders can find it.
 ---
 <p align="center">
-  <strong>Artificial consciousness for machines that work.</strong><br>
+  <strong>Build agents that operate, not just respond.</strong><br />
   <a href="https://apteva.ai">apteva.ai</a>
 </p>

package/package.json CHANGED Viewed

@@ -1,6 +1,6 @@
 {
   "name": "apteva",
-  "version": "0.30.0",
+  "version": "0.30.2",
   "description": "AI agent platform with multi-thread orchestration, MCP integrations, and dashboard",
   "bin": {
     "apteva": "./cli.js"