Kika
Server Details
Shared memory for client engagements: decisions with reasons, blockers, promises, what was tried.
- Status
- Healthy
- Uptime
- 100.0% over the last 24 hours
- Last Tested
- Transport
- Streamable HTTP · MCP 2025-11-25
- URL
- Repository
- usekika/kika-plugin
- GitHub Stars
- 0
- ChatGPT
- –
- Claude
- –
TDQS
Scored across 6 tools
Each tool maps to a distinct resource slice: brief (project status), meeting (metadata/lookup), note (user-typed), summary (AI-generated), transcript (raw utterances), and search_memory (cross-entity retrieval). The main soft spot is get_note vs get_summary, which sound similar until the descriptions clarify human-typed vs AI-generated content, plus the overloading of get_meeting/get_brief as implicit list operations when called without an id.
All six tools are snake_case verb_noun with a single dominant verb family (get_*) plus one search_* for the retrieval tool, which is the natural exception. A caller can predict the shape of any name and what it returns.
Six tools is tightly scoped for a meeting/engagement-context server; each one covers a non-redundant facet (status, meeting, notes, summary, transcript, search) with no filler. Nothing feels missing at the count level and nothing feels bloated.
The read path is well covered end to end: locate a meeting, then pull brief, notes, summary, or transcript, and search across recorded memory. Gaps are around lifecycle and enumeration — no tool to create/append notes or action items, and no first-class list_projects or list_entities (enumeration is folded into get_brief and search_memory's entity parameter), which an agent must discover from prose rather than from the tool surface.
Available Tools
6 toolsget_briefProject briefAInspect
Where an engagement stands: what is blocking it, what you owe and by when, the decisions in force and why, what the client has asked for, and what was already tried and dropped. Call it FIRST, before planning or changing anything — it is the cheapest way to avoid re-deciding something settled or contradicting a call you were not on. Called with no project, it lists the ones this user has.
Inputs: repo, project.
get_meetingFind or open a meetingBInspect
A recorded meeting: when, who spoke, and which parts exist (get_note, get_summary, get_transcript). Without id, finds meetings.
Inputs: id, query.
get_noteMeeting notesCInspect
Notes the user typed in a meeting. Paged.
Inputs: id (required), offset.
get_summaryMeeting AI notesCInspect
AI notes and copilot moments of a meeting. Paged.
Inputs: id (required), offset.
get_transcriptMeeting transcriptBInspect
Timestamped, speaker-attributed transcript of a meeting. Paged.
Inputs: id (required), offset.
search_memorySearch Kika memoryAInspect
Search what this user’s meetings and agents recorded. Pass entity whenever the question names a project, person or company — it anchors the search to that thing rather than to wording, which is more accurate and lets the answer be traced to a source. Call it more than once, broad then narrow.
Inputs: query (required), limit, types, entity, +2 more.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Connectors
Shared team record of decisions, changes and handoffs that every coding agent reads and writes.
Hosted shared memory for AI apps: project decisions, to-dos and learnings. Beta.
Project memory shared by AI agents: decisions, results, tasks and handoffs across sessions.
Governed company memory: search decisions, rules and client facts; propose updates for review.
Related MCP Servers
- AlicenseNot gradedqualityBmaintenanceProvides a shared, repo-local memory that lets coding agents recall relevant decisions, conventions, gotchas, and open threads, save new entries, coordinate file claims, and cache file summaries across tools and sessions.88 npmMIT
- AlicenseNot gradedqualityBmaintenanceEnables a team's AI agents to share one memory of decisions, principles, opinions and open proposals, with authorship, arguments and consensus tracked, so any agent can search what was decided and record new entries or supersede old ones. Contradicting a team decision creates a proposal that another teammate must accept, keeping agreements from being overturned silently.1PolyForm Noncommercial 1.0.0
- AlicenseNot gradedqualityCmaintenanceProvides shared team memory for coding agents, giving each session briefs on teammates' in-progress work, decisions, and assumptions before code is touched, plus handoffs and logged reasons for changes. It lets developers and the wider company stay aligned on what multiple agents are doing without storing any code.MIT
- AlicenseAqualityBmaintenanceEnables AI clients to append durable, content-hashed records of each decision to an owner-controlled markdown notes folder, then recall them across sessions and across different tools with their dates, project provenance, and the exact note versions cited. Records are append-only and searchable by full text, topic stance, or recent activity, so past conclusions can be revisited without relying on an ephemeral session.3317 npmMIT
Glama MCP Gateway
Add one secure layer between your agents and this server.