Skip to main content
Glama

amber_store_memory

Retain text across conversations by storing it as long-term memory. The text is processed into atomic facts, expanded, and embedded; returns a task ID to track and retrieve the resulting memories.

Instructions

Store text as long-term memory. Returns a task_id for tracking. Use amber_get_store_task_status with the task_id to check progress and get the resulting memory IDs when complete. The text is processed in background: chunked into atomic facts, each fact expanded and embedded in parallel, topics resolved (parallel search + sequential creation), then memories inserted in parallel. Typically completes in 10-30 seconds.

Provide metadata when there are specific values worth filtering on later.

TOPICS vs METADATA -- different purposes:

topics = broad subject areas this memory belongs to (like folders). Matched SEMANTICALLY: "career" will find an existing "work" topic automatically, no need to guess exact names. Pick 1-3 areas. Auto-created if no similar topic exists. Examples: "work", "health", "preferences", "family", "cooking", "finances".

metadata = concrete values for EXACT-MATCH filtering in amber_search_memories. Only useful for specific, structured data you want to filter on later: person names, dates, project names, sources. These are NOT semantic -- metadata: {person: "Sarah"} only matches searches filtered by person = "Sarah", not "Sara" or "S. Johnson". Do NOT put categories, importance levels, or topic names in metadata -- that is what topics are for.

Examples: amber_store_memory({ content: "User prefers dark mode in all apps", topics: ["preferences"] })

amber_store_memory({ content: "Meeting with Sarah on Friday to discuss Q3 budget", metadata: {person: "Sarah", date: "2026-05-09", project: "Q3 budget"}, topics: ["work", "meetings"] })

amber_store_memory({ content: "User is allergic to peanuts", topics: ["health"] })

Rate-limited (drip bucket: 1000 capacity, refills ~1 token per 86 seconds. You can burst up to 1000 stores before hitting the limit).

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
contentYesThe text to remember, 1-50000 characters.
metadataNoConcrete values for exact-match filtering in searches. Only useful for structured data like person names, dates, project names, sources. NOT for categories or topic names. Example: {person: "Sarah", date: "2026-05-09"}.
topicsNoOptional broad categories for this memory (e.g. 'work', 'health', 'travel'). Only needed when the category can't be inferred from the text itself -- most memories don't need this. Matched semantically, so you don't need to guess exact names. Do NOT use topics for specific names, dates, or values -- use `metadata` for those (e.g. {"person": "Sarah"}). Topics are searched semantically; metadata is filtered by exact match.
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description fully covers behavioral traits: async background processing, chunking into atomic facts, parallel expansion and embedding, topic resolution, rate limits (drip bucket with capacity and refill rate), and typical completion time.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is well-structured with headers and examples, but it is relatively long. Every sentence adds value, though some redundancy exists (e.g., rate limit details repeated). Front-loaded with key info.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given no output schema, the description explains the return (task_id) and how to check progress to get memory IDs. It covers all parameters, async behavior, rate limiting, and offers examples, making it comprehensive for a complex store tool.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters5/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100% and the description adds significant value beyond the schema by explaining the semantic vs exact-match distinction, providing examples, and clarifying the purpose of each parameter (e.g., topics vs metadata).

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states 'Store text as long-term memory' and explains the async process with a task_id. It distinguishes from siblings like amber_search_memories and amber_get_store_task_status by detailing the background processing and return mechanism.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Provides excellent guidance on when to use metadata vs topics with explicit distinctions and examples. However, it lacks explicit 'when not to use' or direct comparison to other sibling tools beyond the metadata/topics context.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Install Server

Other Tools

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/ambermem/amber'

If you have feedback or need assistance with the MCP directory API, please join our Discord server