Integration surface

44 integration points, named and verifiable.

Ikigai is not a single model with a wrapper. It is a deliberate stack: ten LLMs, eleven IDE / CLI brain-sync targets, twelve media modules, seven voice and image modules, three comms channels, and five infrastructure pillars. Every entry below is in the repo. Every entry is owned by exactly one executive.

LLM providers — 10 first-class

Routed through the Hermes gateway. Each slot is BYOK-capable via /account?tab=mcp.

Claude

claude

Anthropic

opus-4.8 · sonnet-4.6 · haiku-4.5

GPT-5.5

gpt4o

OpenAI

gpt-5.5

Gemini

gemini

Google

gemini-2.5-pro

GLM

glm

Zhipu AI

glm-4.6

Kimi

kimi

Moonshot

kimi-k2.6

Llama

llama

Groq · Meta

llama-4-maverick-17b

DeepSeek

deepseek

DeepSeek

deepseek-v4-pro

Nemotron

nemotron

Nemotron

llama-3.3-nemotron-super-49b

Qwen

qwen

Alibaba DashScope

qwen3-max

Mistral

mistral

Mistral AI

mistral-large-2

Brain-sync targets — 11

Personal Brain Sync. The same Ikigai brain follows the customer into every LLM tool they use.

Claude Code

CLI

Anthropic

.claude/

Cursor

IDE

Cursor

.cursorrules

GitHub Copilot

IDE

GitHub

.github/copilot-instructions.md

Codeium

IDE

Codeium

.codeium/instructions.md

Continue

IDE

Continue

.continue/rules/sid-brain.md

Windsurf

IDE

Codeium

.windsurfrules

Zed

IDE

Zed

.zed/rules.md

Junie

IDE

JetBrains

.junie/guidelines.md

Amazon Q

IDE

AWS

.amazonq/rules/sid-brain.md

Cline

plugin

Cline

.clinerules

Codex / Gemini CLI

CLI

OpenAI · Google

AGENTS.md · GEMINI.md

Media · video modules — 9

The video engine and its sibling modules. Every module is Ikigai-built and Ikigai-owned at the surface level. Wired = active in the render path. Evaluated = available behind a flag.

Ikigai Video Engine

wired

COMPOSITION

Topic → script → scene images → voice-over → music → finished MP4. The end-to-end short-video pipeline.

Ikigai Motion

wired

COMPOSITION

Programmatic brand-reel compositor. Drives the Ikigai Motion render pipeline.

Ikigai Vector Motion

planned

COMPOSITION

Precise vector, data, and voice-synced explainer motion with real-time editor preview.

Ikigai WebGL Stage

planned

COMPOSITION

3D product shots, cinematic camera moves, and interactive premium ad scenes.

Ikigai Codec Layer

wired

COMPOSITION

Static composition + transcode binaries for the render path. Powers every MP4 the stack emits.

Ikigai Graph Composer

evaluated

IMAGE

Node-graph image generation engine. The visual back-end the rest of the image modules render through.

Ikigai Graph Copilot

evaluated

IMAGE

Natural-language copilot over the image graph — describe an image, get a working node graph back.

Ikigai Graph Kit

evaluated

IMAGE

Custom node library + extensions used by the brand reel pipeline.

Ikigai Pixel MCP

evaluated

COMPOSITION

The video engine exposed as a Model-Context-Protocol server so any LLM tool can drive it directly.

Voice · image · composition — 6

ElevenLabs

wired

VOICE

High-fidelity voice synthesis + cloning (paid SaaS)

Edge-TTS

wired

TTS

Free TTS voices, broad language coverage — fallback path

Index-TTS

evaluated

VOICE

Voice cloning module for branded narrator presence

Higgsfield

evaluated

COMPOSITION

Cinematic image-to-video and motion-ad generation (paid MCP)

Flux

wired

IMAGE

Diffusion image generation for thumbnails, posters, scenes

RunningHub

wired

IMAGE

Hosted graph-workflow runner for batch image rendering

Comms channels — 3

wired

9 bot tokens, one per king — per-channel broadcasters

Twilio

wired

Voice + SMS outreach pipe

Google Sheets

wired

Founder + customer webhook export

Infrastructure — 5

Supabase

wired

Postgres + auth + RLS + storage

Vercel

wired

Webapp host + cron + edge runtime

Cloudflare Worker

wired

Edge-fallback layer (workers/edge-fallback.js)

Paystack

wired

ZAR billing + recurring + outcome-share invoices

Resend

wired

Transactional email + outreach delivery

44 integration points, named and verifiable.

LLM providers — 10 first-class

Routed through the Hermes gateway. Each slot is BYOK-capable via /account?tab=mcp.

Claude

claude

Anthropic

opus-4.8 · sonnet-4.6 · haiku-4.5

GPT-5.5

gpt4o

OpenAI

gpt-5.5

Gemini

gemini

Google

gemini-2.5-pro

GLM

glm

Zhipu AI

glm-4.6

Kimi

kimi

Moonshot

kimi-k2.6

Llama

llama

Groq · Meta

llama-4-maverick-17b

DeepSeek

deepseek

DeepSeek

deepseek-v4-pro

Nemotron

nemotron

Nemotron

llama-3.3-nemotron-super-49b

Qwen

qwen

Alibaba DashScope

qwen3-max

Mistral

mistral

Mistral AI

mistral-large-2

Brain-sync targets — 11

Personal Brain Sync. The same Ikigai brain follows the customer into every LLM tool they use.

Claude Code

CLI

Anthropic

.claude/

Cursor

IDE

Cursor

.cursorrules

GitHub Copilot

IDE

GitHub

.github/copilot-instructions.md

Codeium

IDE

Codeium

.codeium/instructions.md

Continue

IDE

Continue

.continue/rules/sid-brain.md

Windsurf

IDE

Codeium

.windsurfrules

Zed

IDE

Zed

.zed/rules.md

Junie

IDE

JetBrains

.junie/guidelines.md

Amazon Q

IDE

AWS

.amazonq/rules/sid-brain.md

Cline

plugin

Cline

.clinerules

Codex / Gemini CLI

CLI

OpenAI · Google

AGENTS.md · GEMINI.md

Media · video modules — 9

The video engine and its sibling modules. Every module is Ikigai-built and Ikigai-owned at the surface level. Wired = active in the render path. Evaluated = available behind a flag.

Ikigai Video Engine

wired

COMPOSITION

Topic → script → scene images → voice-over → music → finished MP4. The end-to-end short-video pipeline.

Ikigai Motion

wired

COMPOSITION

Programmatic brand-reel compositor. Drives the Ikigai Motion render pipeline.

Ikigai Vector Motion

planned

COMPOSITION

Precise vector, data, and voice-synced explainer motion with real-time editor preview.

Ikigai WebGL Stage

planned

COMPOSITION

3D product shots, cinematic camera moves, and interactive premium ad scenes.

Ikigai Codec Layer

wired

COMPOSITION

Static composition + transcode binaries for the render path. Powers every MP4 the stack emits.

Ikigai Graph Composer

evaluated

IMAGE

Node-graph image generation engine. The visual back-end the rest of the image modules render through.

Ikigai Graph Copilot

evaluated

IMAGE

Natural-language copilot over the image graph — describe an image, get a working node graph back.

Ikigai Graph Kit

evaluated

IMAGE

Custom node library + extensions used by the brand reel pipeline.

Ikigai Pixel MCP

evaluated

COMPOSITION

The video engine exposed as a Model-Context-Protocol server so any LLM tool can drive it directly.

Voice · image · composition — 6

ElevenLabs

wired

VOICE

High-fidelity voice synthesis + cloning (paid SaaS)

Edge-TTS

wired

TTS

Free TTS voices, broad language coverage — fallback path

Index-TTS

evaluated

VOICE

Voice cloning module for branded narrator presence

Higgsfield

evaluated

COMPOSITION

Cinematic image-to-video and motion-ad generation (paid MCP)

Flux

wired

IMAGE

Diffusion image generation for thumbnails, posters, scenes

RunningHub

wired

IMAGE

Hosted graph-workflow runner for batch image rendering

LLM providers — 10 first-class

Claude

GPT-5.5

Gemini

GLM

Kimi

Llama

DeepSeek

Nemotron

Qwen

Mistral

Brain-sync targets — 11

Claude Code

Cursor

GitHub Copilot

Codeium

Continue

Windsurf

Zed

Junie

Amazon Q

Cline

Codex / Gemini CLI

Media · video modules — 9

Ikigai Video Engine

Ikigai Motion

Ikigai Vector Motion

Ikigai WebGL Stage

Ikigai Codec Layer

Ikigai Graph Composer

Ikigai Graph Copilot

Ikigai Graph Kit

Ikigai Pixel MCP

Voice · image · composition — 6

ElevenLabs

Edge-TTS

Index-TTS

Higgsfield

Flux

RunningHub

Comms channels — 3

Telegram

Twilio

Google Sheets

Infrastructure — 5

Supabase

Vercel

Cloudflare Worker

Paystack

Resend

LLM providers — 10 first-class

Claude

GPT-5.5

Gemini

GLM

Kimi

Llama

DeepSeek

Nemotron

Qwen

Mistral

Brain-sync targets — 11

Claude Code

Cursor

GitHub Copilot

Codeium

Continue

Windsurf

Zed

Junie

Amazon Q

Cline

Codex / Gemini CLI

Media · video modules — 9

Ikigai Video Engine

Ikigai Motion

Ikigai Vector Motion

Ikigai WebGL Stage

Ikigai Codec Layer

Ikigai Graph Composer