mcp-aurekai
OfficialThe Aurekai MCP server exposes 89+ tools across 9 capability families, giving AI assistants access to media processing, inference, pipelines, knowledge management, commerce, and more — via stdio or HTTP.
Runtime: Async HTTP gateway (
akai_api), auth/sessions (akai_auth), capability discovery (akai_capability), OpenAI-compatible proxy (akai_proxy), CLI dispatch (akai_cli), and change watching (akai_watch).Commerce: Access gates (
akai_gate), payments/invoices (akai_pay,akai_ledger), usage metering (akai_meter), service pricing (akai_finance), and SLA tiers (akai_tier).Intake: Transcribe audio (
akai_transcribe), ingest artifacts (akai_ingest), extract frames (akai_frame_extract), detect scenes (akai_scene_detect), demux video (akai_video_demux), and discover clips (akai_clips).Memory: FPQ embeddings (
akai_fpq,akai_fpqx), vector search (akai_vec), model weight quantization (akai_quant), SAE feature dictionaries (akai_sae), and structured layer inference (akai_sli).Proof: Cryptographic proofs (
akai_proof), data canonicalization (akai_canon), Merkle-DAG graphs (akai_graph), and content hashing (akai_hash).Reason: Physics simulations (
akai_physics,akai_leapfrog), coroutine-native pipelines (akai_flow), inference control (akai_control), and feedback/learning (akai_learn).Wire: Telephony via FreeSWITCH (
akai_tel), QUIC media relay (akai_moq), netlist runtimes (akai_net), and wire captures (akai_wire).Publish: Briefs (
akai_brief), TTS narration (akai_narrate), artifact packing (akai_pack), distribution (akai_distribute), rendering (akai_render), social repurposing (akai_repurpose), and P2P via BitTorrent v2 (akai_swarm).Substrate: Content tagging (
akai_tag), compression (akai_compress), artifact indexing/search (akai_index), entity resolution (akai_entity), document fragmentation (akai_fragment,akai_paragraph), and analytics queries (akai_query).
Additionally, the server provides:
13
aurekai://resources for live data reads (queue stats, models, runtime data, etc.)8 named prompt workflows for common multi-step tasks (e.g., audio → transcribe → brief → deliverable, Merkle lineage inspection, invoice generation, memory pack creation)
Advanced MCP features: tool annotations (
readOnlyHint,destructiveHint,idempotentHint), resource pagination, resource subscriptions,_metaproof propagation, and embedded resource outputs.
@aurekai/mcp — Aurekai MCP Server
0.8.0-alpha.5 · capability-native · zero dependencies · stdio + Streamable HTTP
Exposes all 9 Aurekai capability families (111 commands) as MCP tools with full protocol-level features:
tool annotations, resource pagination, named prompts, _meta proof propagation, and embedded resource outputs.
Install
npm install -g @aurekai/mcpRelated MCP server: CodexPro Runtime
Usage
stdio (default — for Claude Desktop, Cursor, etc.)
// claude_desktop_config.json
{
"mcpServers": {
"aurekai": {
"command": "aurekai-mcp"
}
}
}Streamable HTTP (optional)
AKAI_MCP_HTTP_PORT=3100 aurekai-mcp
# POST JSON-RPC to http://127.0.0.1:3100/mcpProtocol Surface
Feature | Status |
| ✅ |
Tool annotations ( | ✅ |
| ✅ |
| ✅ |
Resource pagination ( | ✅ |
Resource subscriptions (acknowledge) | ✅ |
| ✅ |
| ✅ |
Embedded resource outputs for proof-emitting tools | ✅ |
| ✅ |
Streamable HTTP transport ( | ✅ |
Capability Families
Family | Operators | Examples |
| 11 |
|
| 11 |
|
| 12 |
|
| 11 |
|
| 8 |
|
| 5 |
|
| 5 |
|
| 9 |
|
| 17 |
|
Named Prompts
Prompt | Description |
| audio → transcribe → brief → deliverable |
| Resolve full Merkle lineage for an artifact |
| FPQ compress + roundtrip + export memory pack |
| Dual branch diff with recommendation |
| Metering records → invoice |
| PCAP → SIP event + device report |
| proof validate + manifest verify + SLI auto-run |
| audio → transcript → structured client brief |
Resources (aurekai:// URIs)
aurekai://runtime/capabilities · aurekai://queue/stats · aurekai://ledger/portfolio
aurekai://models · aurekai://model-memory · aurekai://features/{artifact}
aurekai://proof/{id} · aurekai://graph/{node}/lineage · aurekai://space/{name}
aurekai://wire/{capture_id} · aurekai://project/{id} · aurekai://invoice/{id} · aurekai://cms/{entry_id}
Runtime Requirement
Tools require the akai binary on PATH (from aurekai/native-runtime)
or set AKAI_BIN=/path/to/akai. Without it, tools return a clear error message — no crash.
Registry Targets
Available Tools
89 toolsakai_apiC
AkaiAPI v2 — Async HTTP gateway for the Akai binary family. (category: platform)
Inputs: args, stdin.
akai_authC
AkaiAuth — user management and session tokens. (category: platform)
Inputs: args, stdin.
akai_briefD
preserve document order */. (category: artifacts)
Inputs: args, stdin.
akai_canonC
AkaiCanon — structure-aware canonicalization via Tree-sitter. (category: data)
Inputs: args, stdin.
akai_capabilityC
akai-capability - capability discovery and matching layer. (category: platform)
Inputs: args, stdin.
akai_cliC
akai — unified CLI dispatcher. (category: platform)
Inputs: args, stdin.
akai_clipsD
AkaiClips — Clip Discovery Engine. (category: media)
Inputs: args, stdin.
akai_cmsD
AkaiCMS — stripped binary CMS. (category: data)
Inputs: args, stdin.
akai_competeC
akai-compete — stage/model A/B competition engine. (category: specialized)
Inputs: args, stdin.
akai_compressC
AkaiCompress — family-aware compression engine. (category: data)
Inputs: args, stdin.
akai_controlC
akai-control — control plane for Akai inference pipelines. (category: misc)
Inputs: args, stdin.
akai_detect_objectsC
AkaiDetectObjects — Object detection operator for V1/I1 pipelines. (category: specialized)
Inputs: args, stdin.
akai_discipD
AkaiDiscipl operator. (category: specialized)
Inputs: args, stdin.
akai_distributeD
AkaiDistribute operator. (category: artifacts)
Inputs: args, stdin.
akai_economyC
akai-economy — cost-aware routing and budget enforcement. (category: specialized)
Inputs: args, stdin.
akai_embedC
AkaiEmbed — text embeddings via ONNX Runtime C API. Produces: .bfvec. (category: inference)
Inputs: args, stdin.
akai_emitC
AkaiEmit — multi-format output engine. (category: artifacts)
Inputs: args, stdin.
akai_entityD
akai-entity — universal identity resolution layer. (category: knowledge)
Inputs: args, stdin.
akai_familyC
akai-family - conceptual family browser for Akai. (category: specialized)
Inputs: args, stdin.
akai_financeC
AkaiFinance — service arbitrage, labor pipeline, and bundle pricing engine. (category: data)
Inputs: args, stdin.
akai_flash_qlaC
akai-flashqla — AkaiGDN Chunked Prefill (native C kernel). (category: inference)
Inputs: args, stdin.
akai_flowD
akai-flow — coroutine-native pipeline programming. (category: pipeline)
Inputs: args, stdin.
akai_fpqD
main.c — akai-fpq CLI. Produces: .bfq. (category: inference)
Inputs: args, stdin.
akai_fpqxC
akai-fpqx — Cross-family FPQ alignment. Accepts: .bfq. Produces: .bfqx. (category: inference)
Inputs: args, stdin.
akai_fragmentC
AkaiFragment — Fragment system CLI. (category: knowledge)
Inputs: args, stdin.
akai_frame_extractD
akai-frame-extract. (category: media)
Inputs: args, stdin.
akai_gateC
AkaiGate — license enforcement + access control. (category: pipeline)
Inputs: args, stdin.
akai_genC
akai-gen — natural language → recipe YAML generator. (category: data)
Inputs: args, stdin.
akai_graphC
AkaiGraph — Merkle-DAG artifact graph engine. Produces: .bfgraph. (category: knowledge)
Inputs: args, stdin.
akai_hashC
AkaiHash — content-addressing + Merkle DAG hashing engine. (category: artifacts)
Inputs: args, stdin.
akai_indexC
AkaiIndex — artifact family indexer and search engine. (category: knowledge)
Inputs: args, stdin.
akai_ingestC
AkaiIngest — universal intake binary. (category: artifacts)
Inputs: args, stdin.
akai_kv_cacheD
akai-kvcache — v8 RLF KV cache compression. (category: knowledge)
Inputs: args, stdin.
akai_layerC
akai-layer — Layer-aware ONNX model inspection and extraction (C port). (category: knowledge)
Inputs: args, stdin.
akai_leapfrogB
akai-leapfrog — Hamiltonian conservation and reversibility test. Accepts: .bfq, .bfvec. (category: inference)
Inputs: args, stdin.
akai_learnC
akai-learn — artifact-level feedback and threshold tuning. (category: inference)
Inputs: args, stdin.
akai_ledgerD
AkaiLedger — value accounting engine. (category: artifacts)
Inputs: args, stdin.
akai_media_prepD
AkaiMediaPrep operator. (category: media)
Inputs: args, stdin.
akai_meterC
AkaiMeter — usage metering + billing events. (category: platform)
Inputs: args, stdin.
akai_mfa_dictD
AkaiMFADict operator. (category: specialized)
Inputs: args, stdin.
akai_modelC
AkaiModel — model dependency manager for akai pipelines. Produces: .bfmodel. (category: inference)
Inputs: args, stdin.
akai_moqC
akai-moq — Media over QUIC style relay for Akai. (category: specialized)
Inputs: args, stdin.
akai_narrateC
AkaiNarrate v3 — Verified Tone-Aware Text-to-Speech Synthesis. (category: media)
Inputs: args, stdin.
akai_netD
akai-net — Akai Netlist Runtime. (category: platform)
Inputs: args, stdin.
akai_offerD
Also try to read proof-summary.json from. (category: artifacts)
Inputs: args, stdin.
akai_orchestrateD
AkaiOrchestrate operator. (category: pipeline)
Inputs: args, stdin.
akai_outreachD
AkaiOutreach — quiet distribution engine. (category: data)
Inputs: args, stdin.
akai_packC
AkaiPack operator. Produces: .bf, .bfa. (category: artifacts)
Inputs: args, stdin.
akai_paragraphD
AkaiParagraph operator. (category: knowledge)
Inputs: args, stdin.
akai_payC
AkaiPay — payment & invoice management. (category: data)
Inputs: args, stdin.
akai_physicsC
akai-physics — Hamiltonian Version Control Protocol (HVCP) CLI. (category: inference)
Inputs: args, stdin.
akai_pipelineC
AkaiPipeline — unified single-process pipeline. (category: pipeline)
Inputs: args, stdin.
akai_projectC
AkaiProject operator. (category: data)
Inputs: args, stdin.
akai_proofC
AkaiProof operator. Accepts: .bf. Produces: .bfproof. (category: artifacts)
Inputs: args, stdin.
akai_proxyC
AkaiProxy — OpenAI-compatible API shim for Akai binaries. (category: platform)
Inputs: args, stdin.
akai_quantC
akai-quant — FPQ v8 Recursive Lattice-Flow weight quantization. (category: data)
Inputs: args, stdin.
akai_queryC
AkaiQuery — local analytics engine over artifacts via DuckDB. (category: knowledge)
Inputs: args, stdin.
akai_queueC
AkaiQueue v2 — SQLite-backed job queue with built-in worker daemon. (category: pipeline)
Inputs: args, stdin.
akai_reasonD
akai-reason — Reason-state operator over the HVCP stack. (category: specialized)
Inputs: args, stdin.
akai_recipeC
AkaiRecipe — Recipe Registry Management. Produces: .bfrecipe. (category: pipeline)
Inputs: args, stdin.
akai_renderD
AkaiRender operator. (category: artifacts)
Inputs: args, stdin.
akai_repurposeC
AkaiRepurpose — Transform brief artifacts into social media formats. (category: media)
Inputs: args, stdin.
akai_runC
AkaiRun — Recipe Executor. (category: pipeline)
Inputs: args, stdin.
akai_runtimeD
Try top-level sibling path: ../SiblingDir/binary */. (category: platform)
Inputs: args, stdin.
akai_saeC
akai-sae — SAE feature dictionary runtime. Produces: .bfsae. (category: inference)
Inputs: args, stdin.
akai_scene_detectD
akai-scene-detect. (category: media)
Inputs: args, stdin.
akai_segmentC
AkaiSegment — Idea Boundary Detection + Segment Graph. (category: specialized)
Inputs: args, stdin.
akai_sliC
akai-sli — Structured Layer Inference. Accepts: .bfqx, .bfsae. (category: inference)
Inputs: args, stdin.
akai_spaceD
akai-space — shared memory substrate. (category: knowledge)
Inputs: args, stdin.
akai_speech_loopC
AkaiSpeechLoop — Whisper → Transform → Piper speech transformation. (category: media)
Inputs: args, stdin.
akai_stitchC
AkaiStitch — DAG materializer. (category: media)
Inputs: args, stdin.
akai_surfaceD
AkaiSurface operator. (category: specialized)
Inputs: args, stdin.
akai_swarmC
AkaiSwarm — P2P artifact distribution via BitTorrent v2 protocol. (category: pipeline)
Inputs: args, stdin.
akai_syncD
AkaiSync operator. (category: artifacts)
Inputs: args, stdin.
akai_tagC
AkaiTag — instant intent/topic tagging via fastText (pure C). Produces: .bftag. (category: knowledge)
Inputs: args, stdin.
akai_telC
AkaiTel — FreeSWITCH Event Socket telephony adapter. (category: platform)
Inputs: args, stdin.
akai_tierC
akai-tier — latency tier management and SLA enforcement. (category: data)
Inputs: args, stdin.
akai_timeC
akai-time — temporal pipeline manager. (category: data)
Inputs: args, stdin.
akai_toneC
AkaiTone — speech tone/emotion/rhythm extraction via OpenSMILE. Produces: .bftone. (category: media)
Inputs: args, stdin.
akai_transcribeD
================================================================. (category: media)
Inputs: args, stdin.
akai_transcript_cleanD
AkaiTranscriptClean operator. (category: specialized)
Inputs: args, stdin.
akai_transcript_familyD
AkaiTranscriptFamily operator. (category: specialized)
Inputs: args, stdin.
akai_vecC
AkaiVec — local vector search via sqlite-vec. Accepts: .bf. Produces: .bfvec. (category: knowledge)
Inputs: args, stdin.
akai_video_demuxD
akai-video-demux. (category: media)
Inputs: args, stdin.
akai_violenceD
akai-violence — real coupling test harness. (category: specialized)
Inputs: args, stdin.
akai_watchD
AkaiWatch operator. (category: platform)
Inputs: args, stdin.
akai_weaviate_indexD
AkaiWeaviateIndex operator. (category: specialized)
Inputs: args, stdin.
akai_wireC
AkaiWire operator. (category: platform)
Inputs: args, stdin.
akai_workflowC
akai-workflow - workflow profile browser. (category: pipeline)
Inputs: args, stdin.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 89 tools
With 89 tools, many have overlapping purposes despite broad categories. Tools like akai_fpq, akai_fpqx, akai_quant, and akai_flash_qla all relate to quantization/inference, while many 'operator' tools (e.g., akai_discip, akai_fragment) have vague descriptions making them hard to distinguish. The brief and jargon-heavy descriptions further hinder clear differentiation.
All tools share the 'akai_' prefix and use snake_case, but the second part varies between single words (akai_cli, akai_gen) and multi-word phrases (akai_detect_objects, akai_frame_extract). Some are verbs (akai_compress), others nouns (akai_model), showing no consistent verb_noun pattern. This inconsistency makes the set moderately predictable.
89 tools is far above the typical well-scoped range. While the server covers a broad domain (AI/ML pipelines), the sheer number overwhelms an agent's ability to efficiently select the right tool. Many tools appear to be low-level or specialized, suggesting the surface could be consolidated.
Given the large number of tools, the server likely covers many aspects of its domain, but the vague descriptions make it difficult to assess whether CRUD or lifecycle operations are complete. There appear to be many specialized inference and pipeline tools, but gaps in common operations like error handling or logging are apparent.
Maintenance
Related MCP Connectors
Build and run AI workflows, agents, data tables and 1000+ API integrations from any MCP client.
Artifact store for AI agents. Hosted OAuth at mcp.artifacta.io/mcp; local stdio via npm/PyPI.
AI Reasoning Cache & Consensus Layer with 11 MCP tools via Streamable HTTP.
214 tools for AI agents over one MCP endpoint: image, video, audio, 3D, data, storage, mail, GPUs.
Related MCP Servers
- FlicenseCqualityBmaintenanceAn MCP server providing 82 tools for persistent memory, governance, and consciousness continuity for AI instances. Enables local-first, runtime-reflexive operation with cross-device access and recursive self-awareness.843-
- AlicenseBqualityDmaintenanceA local, evidence-driven MCP runtime and control plane for open-source maintainers that provides workspace-bounded tools including controlled file operations, command execution, validation primitives, durable execution records, and human review workflows via stdio and Streamable HTTP transports.33MIT
- AlicenseNot gradedqualityBmaintenanceA cross-platform local AI-agent runtime and MCP gateway that exposes 279 tool definitions (267 advertised by default) for local files, Git, processes, Windows automation, WSL, browser control, Office semantic automation, and extensibility. It enables durable goal continuation, Engineering Harness workflows, context capsules, indexing, and resource-aware concurrent agent tasks through a single governed MCP connection.MIT
- AlicenseBqualityBmaintenanceExposes five prompt-engineering skills as MCP tools, prompts and resources so a connected AI client can itself prepare and execute task, system-prompt, evaluation and rewrite work. It also provides a durable, resumable pipeline that tracks isolated worker dispatches, validates stage output, enforces loop limits and persists run artifacts — locally over stdio or self-hosted on Cloudflare Workers without any model API key.13MIT