GenCMS is iGears' model-agnostic AI workspace and runtime direction for building assistants, document extraction, content and internal agent solutions on an organisation's own data.

It is for developers and organisations that need custom or privately deployed AI with explicit tenant, data and permission boundaries.

Account and tenant administration, workspace records, managed object and vector storage, scoped API keys and Guard scanning are available. Composer can run chat, OCR, embedding, rerank, speech and media generation through a tenant's region-locked AIHK or AICA gateway once its AI key is ready.

Request managed workspace access or ask iGears to scope a 30-day MVP. Project services are fixed-price, with the scope, fee and acceptance checks stated in a quotation.

API

Scoped APIs are live; unified execution is still planned.

Workspace API keys work for tenant-scoped record reads and writes, object and vector storage, Guard scans and Composer runs. Skill execution, Memory retrieval, Workflow runs and Agent triggers are not deployed.

Manage API keys Read API availability

Keys are created, listed and revoked from a signed-in workspace, and the secret is shown once. The available endpoints are listed below; billing, unified credits and a unified Skill, Memory or Agent execution API remain planned.

Concept illustration of a future unified GenCMS execution interface connecting models and runtime modules

Available endpoints

Use these deployed routes today.

GET	/api/v1/workspace	Read workspace counts, policy and recent activity metadata
GET · POST	/api/v1/workspace/skills	List or create Skill install records
GET · POST	/api/v1/workspace/memory	List or create Memory collection records
GET · POST	/api/v1/workspace/memory/{id}/sources	List or create text, URL, file, repository or structured source records
GET · POST	/api/v1/workspace/drafts/{kind}	List or create Agent, Workflow or App drafts
GET	/api/v1/workspace/activity	Read saved workspace activity metadata
GET · PATCH	/api/v1/workspace/settings	Read or update workspace policy
GET	/api/v1/workspace/storage	Read the tenant's object and vector binding status and capabilities
POST	/api/v1/workspace/storage/object/presign	Issue a 60–900 second tenant-scoped S3 GET, PUT or HEAD URL
POST	/api/v1/workspace/storage/vector/query	Query up to 25 points from one tenant-filtered dataset
POST	/api/v1/workspace/storage/vector/upsert	Upsert up to 16 vectors per bounded request
GET	/api/v1/workspace/ai-access	Read the tenant's region assignment, allowed abilities and endpoint examples
POST	/api/v1/guard/scan	Scan a public GitHub repository or supplied text
POST	/api/v1/run	Run Composer through the tenant's assigned AIHK or AICA gateway

read-workspace.sh

# Available with a key carrying workspace:read
curl https://gen-cms.com/api/v1/workspace \
  -H "Authorization: Bearer $GENCMS_API_KEY"

scan-pasted-content.sh

# Available with a key carrying guard:scan
curl https://gen-cms.com/api/v1/guard/scan \
  -H "Authorization: Bearer $GENCMS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "mode": "quick",
    "content": "name: example-skill\npermissions: []"
  }'

Execution roadmap

These endpoint shapes are proposed, not deployed.

They document product direction only and must not be used as a current integration contract.

POST	/v1/skills/{slug}/run	Execute an installed Skill
POST	/v1/memory/{collection}/query	Retrieve indexed context and citations
POST	/v1/workflows/{id}/run	Execute a Workflow draft
POST	/v1/agents/{id}/trigger	Trigger an Agent draft

No unified chat endpoint, Skill runtime, Memory query, Workflow execution, Agent trigger, run billing, streaming contract or usage-credit header is implemented by these proposed routes.

Gateway abilities

One tenant key, six routable abilities.

An administrator assigns each tenant to AIHK in Hong Kong or AICA in Canada and grants the abilities its key may use. Requests never cross into the other region, and a caller cannot select a provider.

chat

OpenAI-compatible chat completions, with vision through data-URI images.

ocr

OpenAI-compatible multimodal requests for reading document images.

embedding

OpenAI-compatible embeddings for the tenant's own indexing.

rerank

Query, documents and top_n reranking.

tts · stt

Text-to-speech synthesis and multipart audio transcription.

media

Queues one asynchronous image or video job and returns a job id, not a finished asset.

Composer itself runs the chat ability only. The other abilities are called directly against the tenant’s assigned gateway with its region-locked key; media requests return a queued job id rather than a finished asset.

Key scopes

Give each integration only the access it needs.

workspace:read

Read tenant-scoped workspace records, settings and activity metadata.

workspace:write

Create, update or remove supported workspace records and settings.

storage:read

Read the storage binding, presign GET and HEAD, and query the tenant's vectors.

storage:write

Presign PUT uploads and upsert vectors into the tenant's dataset.

guard:scan

Run Guard over a public GitHub repository or supplied text.

composer:run

Run chat through the AIHK or AICA gateway assigned to the active tenant.