Sparks AI
A web platform from Lapras Labs for building, running and sharing teams of AI agents. It couples a super-agent, an agent builder, a tool store and real-time chat where humans and agents work side by side inside shared projects.
What is Sparks AI?
Sparks AI is a web platform published by Lapras Labs Inc. that lets one person or a small team assemble, run and share a squad of AI agents. The company presents it as an AI Super App built around AI employees rather than around a single chatbot, and the product is organised in four blocks.
The first is Ion, a super-agent advertised as able to research, code and run tasks autonomously. Its listed capabilities are deep research, instant code generation, agent teams working on autopilot, and the review and editing of artifacts. The second is the agent builder: users start from a pre-built template, pick a language model, then equip the agent with tools and knowledge. Visuals and the older marketing page point to GPT-5, GPT-4, GPT-3.5, Claude Opus, Sonnet and Haiku, and Deepseek. The third block is collaboration — group chat mixing agents and humans, organised into projects with real-time sync. The fourth is the store, where ready-made agents can be browsed and installed, or published for others.
Every paid plan carries the same functional base: tool store access, background agents, web search, image generation, video generation, an agent computer and artifacts. That agent computer rests on third-party sandboxes, judging by the OpenAPI specifications the site publishes for Daytona, E2B and Fly. The same page lists the integrations agents can call: Gmail, Google Calendar, Docs, Drive, Sheets, Slides and Meet, plus GitHub, Notion, Figma, Linear, Shopify, Twitter, YouTube, Cloudflare and Datadog, alongside an MCP entry.
The application itself exposes chats, tasks, artifacts, projects, invitations, shared chats and user profiles. Sign-up is free and can go through Google. A separate enterprise offer, pitched as intelligent automation for cost reduction and operational efficiency, is handled by sales rather than self-service. Pricing runs on monthly credits, from a permanent free tier up to USD 200 per month, with credit packs sold separately.
What it does
- Run Ion, the super-agent, on deep research, instant code generation and autonomous task execution
- Build an agent from a template, pick its language model, then equip it with tools and knowledge
- Connect third-party applications to an agent from the built-in tool store
- Hold group conversations mixing several agents and several humans, synced to shared projects
- Launch background agents and put agent teams on autopilot
- Generate images and video, and produce or edit artifacts such as documents and code
- Publish your own agents to the store, or install ready-made ones from the community
When to use Sparks AI / When not to
A quick filter to help you decide if Sparks AI is the right fit.
When to use Sparks AI
- Small teams and independents who want to hand repetitive research, writing and coding tasks to configurable AI agents
- Developers and technical users who need code generation, sandboxed execution and GitHub, Datadog or Cloudflare connections in one place
- Google Workspace users, since Gmail, Calendar, Docs, Drive, Sheets, Slides and Meet all appear among the published tool specifications
- Product, design and operations people already working in Notion, Figma, Linear or Shopify who want agents acting on those tools
- Buyers on a modest budget who want several model families, web search, image and video generation under a single subscription starting at USD 10 per month
When not to use Sparks AI
- Organisations that require a signed data processing agreement, a published subprocessor list or security certifications such as SOC 2, ISO 27001 or HIPAA, none of which the site publishes
- Teams that must know where their data is stored, since no hosting country or region is named anywhere on the site
- Buyers who need to identify the vendor properly, as no postal address, country of establishment or about page exists
- Engineering teams looking for a public API to embed the product in their own systems, which Sparks AI does not document
- Anyone under 13, excluded by the privacy policy, and anyone needing a native iOS or Android application
How to use Sparks AI
A typical end-to-end flow, from setup to results.
- Open an account from the homepage; sign-up is free and can be done with a Google account
- Pick a pre-built agent template, or start from the agent builder
- Choose the language model that will drive the agent
- Equip the agent with the tools it needs, drawn from the built-in tool store
- Add the knowledge the agent should work from
- Connect the third-party services you already use, such as Google Workspace, GitHub, Notion or Linear
- Create a project so that conversations, tasks and artifacts stay grouped
- Invite colleagues into the project and chat with them and the agents in the same thread
- Delegate longer jobs to Ion or to background agents, then review and edit the artifacts they return
- Track credit consumption and move to a paid plan or a credit pack once the free allowance runs out
Pros & Cons
Pros
- Fully public and figure-by-figure pricing, from a permanent free tier up to USD 200 per month
- Monthly credit allowances stated plan by plan, with separately sold top-up packs
- Wide set of named integrations, published as OpenAPI specifications rather than vague logos
- Humans and agents share the same workspace, organised by projects instead of isolated chats
- Several model families reachable under a single subscription
- Fourteen-day refund on premium features and a stated 99 percent uptime commitment in the terms
- Fifteen percent discount on annual billing, and bonus credits on the larger packs
Cons
- The legal documents are visibly unfinished: the terms still contain drafting instructions and an unreplaced placeholder link to a contact page
- The privacy policy dates from 20 September 2023 and still carries a template placeholder where the company's home country should be
- No postal address, no country of establishment and no about page, so the vendor cannot be properly identified
- No data processing agreement, no subprocessor list and no security certification is published
- Nothing is said about whether customer data trains the models, and no opt-out is documented
- The YouTube link in the footer and the enterprise booking link both returned errors when checked
- Prices and content only exist once JavaScript runs, and the sitemap is broken, which hurts discoverability
Pricing & Plans
Sparks AI offers a permanent free plan, presented as Free Forever, alongside paid subscriptions. The lowest paid entry point is the Lite plan at USD 10.00 per month, billed monthly; annual billing is advertised at fifteen percent less. Credit packs may also be purchased separately, from USD 10.00. Payments are processed through Stripe, and the terms provide for a refund within fourteen days of purchase.
- USD 0 - permanent free tier
- the current pricing page states no figures for it
- USD 10 per month or USD 102 per year - 1
- 000 credits per month - standard processing speed
- email support
- USD 20 per month or USD 204 per year - 2
- 000 credits per month - priority processing
- chat support
- USD 50 per month or USD 510 per year - 5
- 000 credits per month - fastest processing
- priority chat support
- flagged as Most Popular
- USD 200 per month or USD 2
- 040 per year - 20
- 000 credits per month - unlimited processing
- dedicated support
- All paid plans include tool store access
- background agents
- web search
- image generation
- video generation
- agent computer and artifacts
- Starter USD 10 for 1
- 000 credits
- Boost USD 25 for 2
- 700 credits including 200 bonus
- Power USD 50 for 5
- 500 credits including 500 bonus
- Mega USD 100 for 12
- 000 credits including 2
- no published price
- handled through a sales conversation
Data, GDPR & hosting
A consolidated view of how Sparks AI handles your data.
GDPR overview
The privacy policy devotes a section to data protection rights under the GDPR for users in the European Economic Area, and states that Lapras Labs Inc. aims to take reasonable steps to let them correct, amend, delete or limit the use of their personal data. Elsewhere it announces rights of access, rectification, erasure and withdrawal of consent, and a breach notification within an unspecified period. Beyond that, the implementation is thin: no Article 27 EU representative, no data protection officer, no legal basis for processing, no data processing agreement and no subprocessor list. The only privacy contact given is support@askanton.co, an address on a different domain. The policy has not been updated since 20 September 2023.
Who owns the data?
Under the terms and conditions, users keep the copyright to any content they submit. By submitting it, they grant Lapras Labs Inc. a non-exclusive, royalty-free licence to use, distribute, reproduce and display that content in connection with Sparks AI. Feedback and suggestions go further: the company may use, modify and incorporate them with no obligation to compensate the author. On the personal-data side, the privacy policy states that Lapras Labs does not sell or trade user data, but reserves the right to share aggregated, non-identifying demographic information with business partners, and to transfer personal data in a merger, acquisition or asset sale, with prior notice.
Reuse rights
Nothing in the terms restricts what users may do with their own submitted content or with what the agents produce for them: the copyright stays with the user, and no clause requires the vendor's permission for reuse. The counterpart is the broad licence granted to Lapras Labs Inc. over that same content, and an unusually wide feedback clause. Note that the documents never address ownership of AI-generated outputs explicitly, so anything beyond user-submitted material is untested ground. On the vendor side, personal data, usage data and chat history serve personalisation, authentication, transaction processing, customer service and periodic emails about the product or affiliated offers. Google Ads is used for advertising, and third-party providers handle payments and usage analytics.
Data retention & training
Hosting summary
The site names no hosting country and no hosting region. The privacy policy warns that personal data may be transferred outside the user's own jurisdiction, to what it calls the company's home country, but the placeholder for that country was never filled in, so the destination is simply unknown. No trust or security page exists, and no certification such as SOC 2, ISO 27001 or HIPAA is claimed. Technically, the domain resolves to an address announced by Amazon in the United States, and user files are served from a Supabase instance on a Sparks subdomain; neither tells us where data is actually stored, since the resolved address is an anycast node. Anyone with a data residency requirement should treat this as unanswered and ask the vendor directly before committing.
Things to keep in mind
Risks and trade-offs to weigh before adopting Sparks AI.
- Delegating research and writing to autonomous agents erodes the habit of checking sources; the terms themselves warn that outputs may be inaccurate
- The vendor cannot be located: with no address and no country of establishment, there is no practical route to enforce anything against it
- The legal documents are unfinished, so the real scope of the commitments a user accepts is uncertain
- Nothing states whether conversations feed model training, and no opt-out exists, so sensitive material should stay out of the platform
- Chat history is kept for an unspecified duration, and the vendor explicitly declines responsibility for data loss
- Agents that act on connected accounts such as Gmail, Drive, GitHub or Shopify widen the blast radius of a mistaken or manipulated instruction
- Credit-based pricing makes spending hard to anticipate, and heavy autonomous use can consume an allowance quickly
Setup & Integrations
Technical difficulty
Getting started is easy. Sign-up is online, free and possible through Google, with nothing to install; the homepage claims about thirty seconds. Creating a first agent is guided by templates and requires no code. Difficulty rises only when agents are wired to external services, since integrations are described through OpenAPI specifications. No public setup or configuration documentation exists outside the signed-in area, so anything beyond the guided path has to be worked out inside the product.
Deployment
Integrations
Behind Sparks AI
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
What is Sparks AI?
Is there a free plan?
How much do the paid plans cost?
How does billing work?
Which AI models can I use?
What can the agents connect to?
Is there a mobile app or a public API?
What is the minimum age?
Is my data used to train the models?
How do I contact the company?
Should you pick Sparks AI?
Sparks AI is a genuinely broad product for the money. For USD 10 a month a user gets configurable agents, several model families, web search, image and video generation, execution sandboxes and a long list of named integrations, all inside a workspace where colleagues and agents share the same conversation. The pricing is unusually transparent for this segment: every tier, every credit allowance and every top-up pack is stated in figures, with a fourteen-day refund and a stated uptime commitment behind it.
The weakness is not the product but everything around it. The terms and conditions still carry the drafting instructions of the template they came from, and an unreplaced placeholder where a contact link belongs. The privacy policy has not been revised since September 2023 and announces transfers to a home country it never names. There is no postal address, no country of establishment, no about page, no data processing agreement, no subprocessor list and no certification. The only privacy contact published sits on a different domain. Two links advertised on the site were dead when checked, and the sitemap is broken end to end.
None of that says the service works badly, and small teams may reasonably decide it does not matter to them. It does mean an organisation with procurement or compliance obligations has almost nothing to hand its legal reviewers. Sparks AI suits autonomous teams and independents who want capability per euro; buyers in regulated environments should wait for the vendor's paperwork to catch up with its product.
- Choosing a selection results in a full page refresh.
- Opens in a new window.