dakera-mcp
OfficialServer Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| DAKERA_API_KEY | Yes | Your Dakera API key. | |
| DAKERA_API_URL | Yes | The URL of the Dakera API server. | |
| DAKERA_MCP_PROFILE | No | The profile controls which tools appear in tools/list (core. admin, power or all). |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| dakera_storeC | Persist a memory (fact, decision, context) for an agent. importance defaults to 0.5; use 0.8-1.0 for memories that must survive decay. |
| dakera_recallA | Semantic recall: the top_k (default 5) memories closest to a query, optionally narrowed by tags, memory_type or session_id. Session neighbours go in context_memories, outside top_k. include_associated adds KG neighbours one hop away (dakera_recall_associated goes deeper). Best first: rank_score (v0.12.3+) orders them; score may rise. |
| dakera_forgetA | Permanently delete memories by ID or tag. Provide memory_ids for exact removal or tags to bulk-delete all memories sharing those tags. Deletion is immediate and irreversible — prefer dakera_memory_importance to suppress without deleting. |
| dakera_batch_recallA | List memories by filters (tags, importance, time window, type, session) without a query; at most limit (default 100). Use instead of dakera_recall when no semantic search is needed. |
| dakera_batch_forgetA | Bulk-delete memories matching filters (tags, importance range, time window, type, session); at least one is required. Permanent: to demote instead, use dakera_memory_importance. |
| dakera_searchA | Semantic search with optional tag and memory-type pre-filters. Prefer over dakera_recall when results must be constrained by tag or type alongside the semantic match. |
| dakera_session_startA | Open a session; its session_id groups the memories you store. The server ends a session idle for 4 h by default (v0.12.2+). |
| dakera_session_endA | Close a session with an optional summary. Call at run end, even on error. A session the server already ended (idle) keeps its state; your summary is then not saved. |
| dakera_knowledge_graphA | Build a knowledge graph from a seed memory using embedding similarity. Use to explore how a concept connects to stored knowledge. For BFS traversal of an existing linked graph use dakera_graph_traverse. |
| dakera_fulltext_searchA | BM25 keyword search over indexed documents. Use over vector search when exact-term recall matters (error codes, IDs, names). For semantic+keyword combined use dakera_hybrid_search. |
| dakera_hybrid_searchA | BM25 + vector ANN hybrid search in a single pass. Omit vector for BM25-only mode. Use for RAG when pure semantic or keyword search alone is insufficient. vector_weight: 0.0=BM25, 1.0=vector (default 0.5). |
| dakera_extractA | Extract entities, topics, key phrases and a summary from text with the provider chain: extractor_override, then the namespace default, then the server default (GLiNER local). For ad-hoc GLiNER types use dakera_auto_tag. Needs a write key for the namespace (for all namespaces when neither namespace nor agent_id is given). |
| dakera_discover_toolsA | Search the Dakera tool catalog by keyword or tier (core/power/admin/meta) and return names and one-line summaries without loading full schemas. Call this first to find relevant tools, then use dakera_load_tools to fetch only the schemas you need — avoids loading the full catalog upfront. |
| dakera_load_toolsA | Fetch the full inputSchema for one or more named tools. Use after dakera_discover_tools. Returns schemas for found tools and a not_found list for unrecognized names. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 14 tools
The memory-retrieval surface is crowded: dakera_recall, dakera_search, dakera_fulltext_search, dakera_hybrid_search and dakera_batch_recall all return memories, and recall vs search in particular differ only in framing (session neighbours, pre-filters) that an agent must read carefully to distinguish. Store/forget, forget/batch_forget, session_start/session_end, and discover_tools/load_tools are otherwise clearly separated, and the descriptions do actively steer selection.
All names share the dakera_ prefix and snake_case, with verb-first forms dominating (store, forget, recall, search, extract, load_tools). Minor deviations are the noun-only dakera_knowledge_graph and the noun-first session_start/session_end forms, but the set remains highly predictable.
14 tools is a reasonable scope for a memory/KG server, covering write, delete, retrieval, sessions, graph exploration, extraction and catalog meta-tools. Slightly heavy on the retrieval side, where five near-parallel search tools inflate the count, but nothing feels gratuitous.
Multiple descriptions point agents at tools that are not in this set — dakera_memory_importance (suggested repeatedly as the non-destructive alternative to forgetting), dakera_graph_traverse and dakera_auto_tag — creating guaranteed dead ends. There is also no update/edit or importance-adjust operation, so stored memories can only be created or destroyed, not amended.