Skip to main content

​
Platform Tools

Platform tools are pre-built capabilities you can wire up to any agent — managed or custom. No code required. Just enable them and your agent can use them.

Web

Search & FetchSearch the web and fetch content from URLs. Get real-time information for your agents.

Memory

User-Scoped PersistenceRemember information across conversations. Each user gets their own isolated memory.

Knowledge Base

RAG Over Your DocsUpload documents and let agents query them semantically. Project-scoped.

KV Storage

Lightweight StateKey-value storage for counters, preferences, flags, and simple state.

​
Platform Tools vs Custom Tools

Platform ToolsCustom Tools
SetupEnable in UI, no codeWrite Python or TypeScript
HostingManaged by ReminixDeployed with your code
Use caseCommon capabilitiesDomain-specific logic
ExamplesWeb search, memory, knowledgeCRM lookup, payment processing
Both types of tools can be used together. A managed agent can use platform tools, and a custom agent can use both platform and custom tools.

​
Web Tool

Search the web and fetch content from URLs.

​
Capabilities

  • Search: Query search engines for real-time information
  • Fetch: Retrieve and parse content from any URL

​
Use Cases

  • Answer questions about current events
  • Research topics in real-time
  • Extract data from web pages
  • Verify facts and information

​
Example

When enabled, your agent can answer questions like:
  • “What’s the latest news about AI?”
  • “What does the documentation at this URL say?”
  • “Find me the current stock price of AAPL”

​
Memory Tool

User-scoped persistence that remembers information across conversations.

​
How It Works

Memory is scoped to individual users. When a user interacts with your agent, the agent can:
  1. Store information about the user
  2. Retrieve previously stored information
  3. Update existing memories
  4. Clear memories when needed

​
Use Cases

  • Remember user preferences
  • Track conversation context across sessions
  • Build personalized experiences
  • Maintain user-specific state

​
Example

User: My name is Alex and I prefer dark mode.
Agent: Got it, Alex! I'll remember your dark mode preference.

[Later session]
User: What are my preferences?
Agent: You prefer dark mode, Alex.

​
Scoping

Memory is user-scoped by default. Each user’s memory is isolated — Agent can’t access User A’s memories when talking to User B.

​
Knowledge Base Tool

RAG (Retrieval-Augmented Generation) over your uploaded documents.

​
How It Works

  1. Upload documents to your project (PDFs, text files, markdown)
  2. Enable the knowledge base tool for your agent
  3. Query — the agent automatically searches relevant documents when answering questions

​
Use Cases

  • Answer questions about your product documentation
  • Build support agents that know your help articles
  • Create research assistants over large document sets
  • Power FAQ bots with accurate, sourced answers

​
Scoping

Knowledge base is project-scoped. All agents in a project share the same knowledge base.

​
Supported Formats

  • PDF
  • Markdown (.md)
  • Plain text (.txt)
  • HTML

​
KV Storage Tool

Lightweight key-value storage for simple state.

​
How It Works

Store and retrieve key-value pairs. Useful for:
  • Counters (e.g., “How many times has this user asked about pricing?”)
  • Flags (e.g., “Has onboarding been completed?”)
  • Simple preferences
  • Session state

​
Example

# Agent stores a value
kv.set("user:123:onboarding_complete", True)

# Agent retrieves it later
is_complete = kv.get("user:123:onboarding_complete")  # True

​
When to Use KV vs Memory

Use CaseTool
Remember what user said in conversationMemory
Track a counter or flagKV Storage
Store user preferences (natural language)Memory
Store user preferences (structured)KV Storage

​
Enabling Platform Tools

​
For Managed Agents

  1. Go to your agent in the Dashboard
  2. Navigate to the Tools tab
  3. Toggle on the platform tools you want to enable
  4. Save

​
For Custom Agents

Custom agents can call platform tools via the Reminix SDK when running on Reminix Cloud: 
from reminix_runtime import agent, platform_tools

@agent
async def my_agent(prompt: str) -> str:
    # Search the web
    results = await platform_tools.web.search(prompt)
    
    # Store in memory
    await platform_tools.memory.set("last_search", prompt)
    
    # Query knowledge base
    docs = await platform_tools.knowledge.query(prompt)
    
    return f"Found {len(results)} results and {len(docs)} relevant documents"
Platform tools are only available when running on Reminix Cloud. When self-hosting, you’ll need to implement equivalent functionality yourself or use custom tools.

​
Limits by Plan

Platform tool usage varies by plan:
ToolFreeProTeam
Web searches50/month1,000/month10,000/month
Memory operations100/month10,000/monthUnlimited
Knowledge base storage10 MB100 MB1 GB
KV operations1,000/month50,000/monthUnlimited
See Pricing for full details.

​
Next Steps

Custom Tools

Build your own tools in Python or TypeScript

Agents

Learn about managed and custom agents