Boost.space AgentOS
Boost.space AgentOS runs a team of specialized AI agents on a shared company data layer, with per-agent data and app permissions, reusable skills, approved ACT automations, MCP, a REST API and TypeScript/Python SDKs. Lifetime tiers from $79 on AppSumo; done-for-you agent launches from $2,500.
About Boost.space AgentOS
What Boost.space AgentOS is
AgentOS is a platform for running a team of specialized AI agents inside a company, from Boost.space, the Czech data-synchronization company. Each agent has one business role (following up on leads, enriching a product catalog, writing a weekly report), works on the company's own data, uses the apps and skills it has been allowed, and runs on a schedule as well as on request. People hand out work through a chat assistant called Boosty, by text or voice, which routes each job to the agent whose role fits and brings the result back. AgentOS premiered on November 16, 2026, the same day it launched on Product Hunt as "One operating system for all your AI agents".
The docs are explicit that AgentOS is a separate product from classic Boost.space, the data platform the company launched in September 2022 (its Trust Center gives 2017 as the founding year). The two share core technology but have their own licences, systems and limits, and can be connected through the API or MCP. The company's own showcase agent is Bob, a product-catalog agent for commerce teams that works on GEO visibility in AI answers, dynamic pricing, product enrichment, marketplace listings and retention marketing.
Building blocks
- Agents and instances. An agent is a specialized worker with one job; it takes no user seat. Agents run on an instance, a server that Boost.space cloud runs for you in the EU, or one of your own.
- Data layer. Company data lives in modules, records and fields shared by people and agents, in a separate system per customer at
https://<system>.boost.spacewith its own database, users, keys and MCP endpoint. - Skills and connections. Skills are reusable know-how shared between agents, each carrying a security-scan result. Connections are the apps an agent may use; the docs list 1,649 apps, any app's own MCP server can be added, and 16 messaging channels (Slack, Teams, WhatsApp, Telegram, email and more) let people reach an agent where they chat.
- Automations. An agent automation uses the agent's judgement on every run; an ACT automation is code the agent writes, tests and switches on after your yes, and then runs the same way every time, each run in its own discarded environment.
- Models. The docs recommend your own AI account for daily work: a ChatGPT subscription, a Claude plan, or API keys from OpenAI, Anthropic, Google, xAI, DeepSeek and others, Azure AI Foundry, Amazon Bedrock or an OpenAI-compatible server. The AI wallet (2,000 AI Credits = $1) pays for trying models and for images, video, voice, web data and paid apps.
- Developer access. A REST API, typed TypeScript and Python SDKs, and MCP so ChatGPT, Claude, Claude Code, Cursor or VS Code can work with your AgentOS data and hand tasks to agents under your own permissions.
Admins set access agent by agent: for each data module No access, Read or Read & write; which connected accounts each agent may use; and who may message it in each channel. Before an ACT automation calls another service or goes live, the agent shows what it will do and waits for approval. Secrets are entered through a one-time link valid for 10 minutes and can't be read back by people or agents.
Pricing (checked November 17, 2026)
Boost.space currently sells AgentOS through two routes, and they look very different:
| Route | Price | What you get |
|---|---|---|
| Agent Launch (boost.space) | $2,500 one-time for Bob, $5,000 for a custom role, excl. VAT | One agreed first goal, up to 8 hours of expert work, the delivered agent (training, memory, configuration, skills) to keep, first month of AgentOS included |
| AgentOS managed plan (boost.space) | From $300/month billed annually ($3,600/year) | Entry plan lists 100,000 managed records and 100,000 AI Credits a month; plan confirmed before purchase |
| AppSumo lifetime tiers (docs) | $79 to $3,499 one-time | Tier 1: 3 agents, 1 instance, 10,000 ACT Operations/month, 10,000 records, 1 user. Tier 6: 300 agents, 50 instances, 1,000,000 operations, 2,000,000 records, 200 users. $1 of AI Credits a month on every tier |
AI provider usage, hosting and paid external services are separate unless an agreement includes them. The company states that no positive ROI is guaranteed for the four-week Bob launch.
Worked example: sizing ACT Operations
The FAQ defines one ACT Operation as up to 5 seconds of an automation's run time, counting every started 5-second block (an 11-second run uses 3), with test runs included and chat use excluded. We sized a realistic workload: a catalog price check that runs every 15 minutes and takes about 40 seconds, plus a nightly enrichment job of about 3 minutes.
price check: ceil(40 s / 5) = 8 ops x 96 runs/day x 30 days = 23,040 ops
nightly job: ceil(180 s / 5) = 36 ops x 30 nights = 1,080 ops
total = 24,120 ops/month
AppSumo Tier 1 (10,000) -> too small
AppSumo Tier 2 (20,000) -> too small
AppSumo Tier 3 (50,000) -> fits, ~26k headroom for test runs and growth
Two lessons from the arithmetic: frequency dominates (the 15-minute job is 95% of the total), and rounding up to 5-second blocks punishes short, frequent runs, so batching work into fewer runs saves operations. Waiting counts too, so an automation that sleeps on a slow API burns operations while it waits.
We checked the SDK packages
npm @boost-space/sdk latest 0.1.3 license MIT 3 versions
first published 2026-07-20 engines: node >=20
repository: github.com/boostspace/sdk (packages/typescript)
PyPI boostspace-sdk 0.1.3 requires Python >=3.11 3 releases since 2026-07-20
GitHub org boostspace (public): boostspace-mcp-server, integrator-mcp-server,
mcp-servers, n8n-nodes-boostspace -> no public "sdk" repository (checked 2026-11-17)
The SDKs are real and installable, but they are early (0.1.x) and the repository the npm package points to is not public, so you can't read or file issues against the source. The docs say the SDK retries 429s after Retry-After and retries 5xx errors for GET, PUT and DELETE, but never retries POST, so nothing is created twice.
AgentOS vs the alternatives
| Boost.space AgentOS | n8n / Zapier / Make (see our comparison) | Rakazo | |
|---|---|---|---|
| Core idea | Role-based agents on a shared company data layer | Workflow automation with AI steps | Persistent AI bots with memory and routines |
| Data store | Built-in modules/records per company | Mostly passes data between apps | Bot memory |
| Repeatable runs | ACT automations, billed per 5-second block | Workflows, billed per task/operation/execution | Markdown routines |
| Self-host | Agents on your own server; data system in Boost.space's EU cloud | n8n yes; Zapier and Make no | Yes, Apache-2.0 |
| Done-for-you option | Agent Launch with Boost.space experts | Partners/agencies | No |
For the underlying choice between autonomous agents and fixed workflows, see AI agents vs AI workflows; AgentOS deliberately mixes both.
Who it's for
Small and mid-size companies, especially e-commerce and catalog-heavy businesses, that want agents doing ongoing operational work on their own data; agencies building agents for clients (the top AppSumo tier is designed for up to 35 client platforms, marked coming soon); existing Boost.space customers who want to add agents alongside their data platform; and developers who want MCP and an API over a governed company data layer.
Limits and trade-offs
- Two price lists. The website's done-for-you route (thousands of dollars plus a monthly plan) and the docs' AppSumo lifetime tiers (from $79) describe very different budgets; confirm which applies to you before planning.
- Small AI allowance. Every AppSumo tier includes $1 of AI Credits a month, which the docs say an agent thinking on the wallet "uses up quickly"; plan on connecting your own AI account and paying that provider.
- EU cloud for data. Agents can run on your own server, but your AgentOS system and database are hosted by Boost.space in the EU. Unlimited agents on your own servers (Tiers 5 and 6) and client platforms are marked coming soon.
- Certification level. Boost.space lists SOC 2 Type I (a point-in-time design review, not the longer Type II), ISO/IEC 27001:2022, GDPR, HIPAA and Google CASA Tier 2.
- Classic Boost.space is changing. The Make.com engine built into classic Boost.space will be replaced by the new agentic automation, with timelines promised for December 2026; AgentOS itself is a separate purchase.
- Support hours. Human support replies on workdays, 9 AM to 7 PM CET.
- Vendor claims. "Trusted by 15,000 teams in 140 countries", "9B+ commerce records" and "1,400+ AI models" are Boost.space's figures.
Related on TechLogHub
- What is an AI agent? and agentic AI.
- Model Context Protocol (MCP) and MCP vs REST API, the two ways to reach AgentOS data.
- REST API, SDK and webhooks, the developer surfaces AgentOS exposes.
- Cron expression generator, handy when scheduling agent and ACT automations.
Value & Audience
Value Proposition
Give each AI agent one business role, the company data and apps it needs and nothing more, and let it run that job on a schedule, with repeatable steps turned into tested automations.
Problem Solved
Chat assistants answer questions but don't carry a job forward, and every new agent a company builds has to be wired again to the same data, apps and permissions.
Tech Stack
Social & Links
Loading ratings...
Loading comments...
Gallery

Feedback & Roadmap
Boost.space AgentOS FAQ
Common questions about Boost.space AgentOS's features, pricing, and use cases
Is Boost.space AgentOS free to use?
No. Boost.space AgentOS is a paid product, starting at AppSumo lifetime - $79 to $3,499 one-time - Tier 1: 3 agents, 1 instance, 10,000 ACT Operations/month, 10,000 records, 1 user; $1 AI Credits/month on every tier. Plans listed: AppSumo lifetime - $79 to $3,499 one-time - Tier 1: 3 agents, 1 instance, 10,000 ACT Operations/month, 10,000 records, 1 user; $1 AI Credits/month on every tier; Agent Launch - $2,500 (Bob) or $5,000 (custom role) one-time, excl. VAT - up to 8 hours expert work, delivered agent stays yours, first month of AgentOS; AgentOS managed plan - from $300/month billed annually ($3,600/yr) - 100,000 managed records, 100,000 AI Credits/month.
What are some alternatives to Boost.space AgentOS?
Popular AI Agents & Automation alternatives to Boost.space AgentOS include NeverApply, Composio and Harness².
Boost.space AgentOS appears in
You May Also Like

NeverApply
AI job-search agent that matches jobs, tailors your resume, and auto-applies.
View ProductComposio
Composio connects AI assistants and custom agents to more than 1,500 apps through managed authentication, SDKs and an MCP gateway, and its new Instant tools let agents pay per call for search, scraping, audio and data services from one balance.
View Product
Harness²
Harness² is an Apache-2.0 research framework from Google Cloud AI Research that automatically refines an LLM agent's harness (prompts, tools, memory, skills and more) for each task and carries lessons forward to future tasks, without changing model weights.
View ProductCraft Agents
Craft Agents is an open-source (Apache 2.0) desktop app from the makers of Craft Docs for working with AI agents. It connects agents to MCP servers, REST APIs, local files and a built-in browser, with multi-session workspaces, Skills, Automations and Pages.
View ProductOne email a week
TechLogHub tracks 878 tools and 602 open-source projects. Get the week's new launches, release roundups and open-source picks — nothing else.
Get the TechLogHub digest
One email a week: new and trending developer tools, fresh comparisons, and what shipped. Unsubscribe in one click.