RationalBloks

 ⟨93 unchanged words⟩ "
• DON'T define: id, created_at, updated_at, row_version
• NEVER create users/customers/employees tables (use app_users) ⟨85 unchanged words⟩ 

 ⟨88 unchanged words⟩  ONE_TO_ONE, ONE_TO_MANY, MANY_TO_ONE, MANY_TO_MANY

HIERARCHICAL NODES:
Nest entities inside parent entities to createnest typean hierarchies.
Childentity entitiesinside inheritits parent labels automatically.

Example:
{
  "nodes":entity {
    "Animal":(beside {
      "description": "Base animal entity",
       "flat_labels":
and ["LivingThing"],
      "schema": {
        "name": {"type":) "string",to "required":create true},
        "habitat":a {"type": "string"}
      },
      "Dog": {
        "description":hierarchy; "Athe dogchild (inherits Animalthe labels)",
        "flat_labels":parent's ["Pet"],
        "schema"labels: {
          
  "breedAnimal": {"typedescription": "string...", "required": true},
          "trainedschema": {"type": "boolean"}
        }
      }
    }
  ...},
  "relationships": {
    "OWNSDog": {
      "fromdescription": "Person...",
      "to": "Animal",
      "cardinalityschema": "ONE_TO_MANY"
    {...}
  }
}

RULES:
1. "nodes" key is REQUIRED ⟨55 unchanged words⟩  (id, created_at, updated_at) are NOT needed
8. Use get_graph_template_schemas FIRST to see valid examples

WORKFLOW:
1. Use get_graph_template_schemas FIRST to see valid examples
2. Createcreate_graph_project schemawith followinga thecluster_id rulesfrom abovelist_clusters
3. Call this tool
4. Monitor with get_job_status (2-5 min deployment)

After creation, usePoll get_job_status with the returned job_id to monitor(2-5 deployment.minutes) While RationalBloks is being updated, the call  ⟨14 unchanged words⟩ 

{"destructiveHint":false,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

{"destructiveHint":false,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

Create a new RationalBloks project (a PostgreSQL REST API) from a JSON schema.schema, on one of your resource pools.

⚠️ CRITICAL RULES - READ BEFORE CREATING SCHEMA:

1. FLAT FORMAT (REQUIRED):
   ✅ CORRECT: {users"orders": {email"total": {"type": "stringdecimal", max_length"precision": 25510, "scale": 2}}}
   ❌ WRONG: {users"orders": {"fields: {email": {type: "string"}...}}}
   DO NOT— never nest under 'fields' key!

2. FIELDEVERY TYPEfield REQUIREMENTShas a "type":
   • string: (MUST have "max_length" (e.g.), max_length:text, 255)
   •integer, decimal: MUST have "precision" and "scale" (e.g., precision: 10, scale: 2)
   •, datetime:boolean, Useuuid, "date,
   datetime" (NOT "timestamp"
   •), ALLjson, fieldsand the PostgreSQL arrays uuid_array, integer_array, text_array,
   float_array (GIN-indexed: MUST@> havecontains, "type"<@ propertycontained_by, && overlaps)

3. AUTOMATIC FIELDS (DON'T define):
   • id (uuid, primary key)
   • created_at (datetime)
   •, updated_at (datetime)

4. USER AUTHENTICATIONUSERS:
   ❌ NEVER create "users", "customers", "employees"users/customers/employees tables with email/password
   ✅email USEor built-inpassword. Link to the built-in
   app_users table

   Example:
   {
     "employee_profiles": {
       "user_id": {type: "uuidtype", foreign_key: "app_users.iduuid", required: true},
       "departmentforeign_key": {type: "stringapp_users.id", max_length: 100}
     }
   }

5. AUTHORIZATION:
   Add user_id → app_users.id to enable. "onlyA seetable yourwith ownsuch data"

   Example:a
   {
     "orders": {
       "user_id": {type: "uuid",field foreign_key:is "app_users.id"},
       "total"owner-scoped: {type:each "decimal",user precision:sees 10,their scale:own 2}
     }
   }rows.

6.5. FIELD OPTIONS:
   • required: true/false
   •, unique: true/false
   •, default: any value
   •, enum: ["val1", "val2"]
   • foreign_key: ("table.id"

AVAILABLE TYPES: string, text, integer), decimal,on_delete boolean(cascade, uuid
   set_null, daterestrict, datetimeno_action), jsonnullable, uuid_arrayindex, integer_arraydescription, text_arraywrite_only, float_array

   Array types store PostgreSQL native arrays with automatic GIN indexing:computed.
   • uuid_array: UUID[] — for sets ofAdvanced referencesfeatures (e.g., tensor coordinates)
   • integer_array: BIGINT[] — for dimension indices, integer sets
   • text_array: TEXT[] — for tags, categories__policy__, label sets
   • float_array: DOUBLE PRECISION[] — for weightcomputed vectorscolumns, scores
   GIN-indexed operators: @> (contains)__constraints__, <@ (contained_by__audit__), &&: (overlaps)get_schema_reference.

BACKEND ENGINE:
• python (defaultbackend_type): FastAPI backend —"python" mature(FastAPI, full-featured
•default) or "rust:" (Axum backend —: faster cold starts, lower memory, high performance).

WORKFLOW:
1. Use get_template_schemas FIRST to see valid examples
2. Createcreate_project schemawith followinga ALLcluster_id rulesfrom abovelist_clusters
3. Call this tool (optionally choose backend_type: "python"Poll orget_job_status "rust")
4.with Monitorthe withreturned get_job_statusjob_id (2-5 min deploymentminutes)

After creation, use get_job_status with
4. returnedget_project_info job_idgives tothe monitorlive deployment.URL While RationalBloks is being updated, the call  ⟨14 unchanged words⟩ 

{"destructiveHint":false,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

{"destructiveHint":true,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

{"destructiveHint":true,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

{"destructiveHint":true,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

{"destructiveHint":true,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

 ⟨13 unchanged words⟩ own credentials and database. Requires paid plan. A deployplan that drops data is refused, untilnaming youevery passtable, confirm_destructive=truefield, afterentity reviewingor therelationship plan.it would drop, and every field whose type change rounds or cuts its values (a decimal with a smaller scale or turned integer, a datetime turned date): deploy_destructive applies it, a tool of its own so a client asks before it runs. One operation runs on a project at a ⟨44 unchanged words⟩ 

{"properties":{"confirm_destructive":{"description":"Set true only after reviewing the plan: a deploy that drops tables, columns, entities, relationships or fields is refused without it","type":"boolean"},"project_id":{"description":"Project ID (UUID) ⟨6 unchanged words⟩ 

{"destructiveHint":truefalse,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

Deploy a graph projectproject's saved schema to the staging environment. This triggers: (1) ⟨39 unchanged words⟩ minutes. Use get_graph_project_info to verify deployment succeeded. A deployplan that drops data is refused, untilnaming youevery passtable, confirm_destructive=truefield, afterentity reviewingor therelationship plan.it would drop, and every field whose type change rounds or cuts its values (a decimal with a smaller scale or turned integer, a datetime turned date): deploy_destructive applies it, a tool of its own so a client asks before it runs. One operation runs on a project at a ⟨44 unchanged words⟩ 

{"properties":{"confirm_destructive":{"description":"Set true only after reviewing the plan: a deploy that drops tables, columns, entities, relationships or fields is refused without it","type":"boolean"},"project_id":{"description":"Project ID (UUID) ⟨6 unchanged words⟩ 

{"destructiveHint":truefalse,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

Promote staging to production (requires a paid plan): production's database migrates from the schema it runs to the one staging runs, and production runs staging's build. With staging unchanged since the last promotion, it rebuilds production with no schema change. The operation is ASYNCHRONOUS: poll the returned job_id with get_job_status. A deployplan that drops data is refused, untilnaming youevery passtable, confirm_destructive=truefield, afterentity reviewingor therelationship plan.it would drop, and every field whose type change rounds or cuts its values (a decimal with a smaller scale or turned integer, a datetime turned date): deploy_destructive applies it, a tool of its own so a client asks before it runs. One operation runs on a project at a ⟨44 unchanged words⟩ 

{"properties":{"confirm_destructive":{"description":"Set true only after reviewing the plan: a deploy that drops tables, columns, entities, relationships or fields is refused without it","type":"boolean"},"project_id":{"description":"Project ID (UUID) ⟨6 unchanged words⟩ 

{"destructiveHint":truefalse,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

Deploy a projectproject's saved schema to the staging environment. This triggers: (1) ⟨96 unchanged words⟩ get_project_info to see if the deployment succeeded. A deployplan that drops data is refused, untilnaming youevery passtable, confirm_destructive=truefield, afterentity reviewingor therelationship plan.it would drop, and every field whose type change rounds or cuts its values (a decimal with a smaller scale or turned integer, a datetime turned date): deploy_destructive applies it, a tool of its own so a client asks before it runs. One operation runs on a project at a ⟨44 unchanged words⟩ 

{"properties":{"confirm_destructive":{"description":"Set true only after reviewing the plan: a deploy that drops tables, columns, entities, relationships or fields is refused without it","type":"boolean"},"project_id":{"description":"Project ID (UUID) ⟨6 unchanged words⟩ 

{"destructiveHint":truefalse,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

 ⟨5 unchanged words⟩  job (a create, deploy, promotion, rollback, deletion or deletionmodule operation). STATUS VALUES: pending (queued), processing ⟨30 unchanged words⟩  a deploy can take up to 15 minutes. A module job names its module_code, and a completed module build's result names the image it built (its commit). If status is 'failed', read failure_side and  ⟨54 unchanged words⟩ 

{"properties":{"job_id":{"description":"Job ID returneda fromcreate, deploymentdeploy, operationspromotion, rollback, deletion or module operation answered","type":"string"}},"required": ⟨3 unchanged words⟩ 

Change PART of a project's schema without dropping any of it (saves to database, does NOT deploy).

 ⟨64 unchanged words⟩  rules as create_project
• {"op": "drop_table", "table": "clocks"}
• {"op": "rename_table", "table": "clocks", "to ⟨12 unchanged words⟩ ": 64}}
• {"op": "drop_field", "table": "parameters", "field": "is_clock"}
• {"op": "rename_field", "table": "parameters", "field ⟨28 unchanged words⟩  true}}   __policy__, __constraints__, __audit__, __admin_write__

Dropping a table or field is drop_schema_items, a tool of its own.

ALL OR NOTHING: an operation that cannot be ⟨59 unchanged words⟩ 'version' (the schema's version
after the change). You never need to read the whole schema back to see what you did.

WORKFLOW:
1. get_schema with tables/fields to read  ⟨5 unchanged words⟩ (its 'version' is the whole schema's)
2. patch_schemapreview_schema_change with dry_run=truethe operations: check 'diff' and 'plan'
3. patch_schema with  ⟨7 unchanged words⟩ refused if someone else saved meanwhile)
4. deploy_staging (confirm_destructive=true when the plan drops data), then get_job_status

{"properties":{"dry_run":{"description":"Save nothing: answer what the operations did, the diff and the migration plan a deploy would run.","type":"boolean"},"expected_version":{"description":"The 'version' a get_schema ⟨60 unchanged words⟩ 

{"destructiveHint":false,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

{"destructiveHint":true,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

{"destructiveHint":true,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

{"destructiveHint":false,"idempotentHint":true,"openWorldHint":truefalse,"readOnlyHint":false}

UpdateReplace a graph project's WHOLE schema (saves to database, does NOT deploy). DESTRUCTIVE: an entity, relationship or field the schema leaves out is deleted by the next deploy.

⚠️ Follow ALL rules from create_graph_project:
•  ⟨55 unchanged words⟩  Entity names should be PascalCase

WORKFLOW:
1. Use get_graph_schema to see the current schema
2. Modify following all rules
3. Call update_graph_schema (saves only)
4. Callits deploy_graph_staging'version' topins applythe changessave)
5.2. Monitorpreview_schema_change with get_job_status

DRY RUN: pass dry_run=truethe tonew previewschema: whatcheck a'diff' deployand WOULD'plan'
3. changeupdate_graph_schema (renames, deletions) withoutpass saving.

NOTE:expected_version Thisto onlybe savesrefused theif schema.someone Youelse MUSTsaved callmeanwhile)
4. deploy_graph_staging afterwards, tothen deploy.get_job_status

{"properties":{"dry_runexpected_version":{"description":"PreviewThe the'version' planneda migrationget_graph_schema read answered. The save is refused (renames/deletions409) without savingwhen orthe deploying.saved Nothingschema ischanged applied.since.","type":"booleanstring"},"project_id":{"description":"Project ID ⟨5 unchanged words⟩ 

{"destructiveHint":falsetrue,"idempotentHint":true,"openWorldHint":truefalse,"readOnlyHint":false}

UpdateReplace a project's WHOLE schema (saves to database, does NOT deploy).

⚠️ CRITICALDESTRUCTIVE: a table or field the schema leaves out is dropped by the next deploy.

For a new data model or a template. To change part of a schema use patch_schema, and drop_schema_items to drop; sending back a large schema to change one field risks changing something else on the way.

⚠️ Follow ALL rules from create_project:
• FLAT format ⟨50 unchanged words⟩  add required field without default to existing tables
• Safe: {new_field: {type: "string", max_length: 100, required: false}}

WORKFLOW:
1. Use get_schema to see the current schema
2. Modify following ALL rules
3. (optional) Call update_schema with dry_run=trueits to'version' previewpins the migration first
4. Call update_schema (saves onlysave)
5. Call deploy_staging to apply changes
6.2. Monitorpreview_schema_change with get_job_status

CHANGING PART OF A SCHEMA: use patch_schema. This tool replaces the WHOLE schema, so it is for a
new data model or a template; sending back a large schema to change: onecheck field'diff' risksand changing'plan'
something else on the way.

DRY3. RUN:update_schema (pass dry_run=trueexpected_version to savebe nothingrefused andif readsomeone backelse 'diff'saved —meanwhile)
4. everydeploy_staging, propertythen thisget_job_status

A schema changes
in the saved one, sodeploy anwould accidentallyrefuse changedis description,refused defaulthere, enum or computed case shows up —
and 'plan', thenothing migrationis asaved. deployThe wouldanswer thencarries run'diff', with destructive steps flagged.

NOTE: Without dry_run this only'plan' savesand the schema. You MUST call deploy_staging afterwards to applynew changes.'version'.

{"properties":{"dry_run":{"description":"Save nothing: answer the property-level diff against the saved schema and the migration plan a deploy would run.","type":"boolean"},"expected_version":{"description":"The 'version' a get_schema ⟨36 unchanged words⟩  FLAT format (table_name → field_name → properties). — the WHOLE schema. Every field MUST have a 'type' property.", ⟨7 unchanged words⟩ 

{"destructiveHint":falsetrue,"idempotentHint":true,"openWorldHint":truefalse,"readOnlyHint":false}

 ⟨88 unchanged words⟩  ONE_TO_ONE, ONE_TO_MANY, MANY_TO_ONE, MANY_TO_MANY

HIERARCHICAL NODES:
Nest entities inside parent entities to createnest typean hierarchies.
Childentity entitiesinside inheritits parent labels automatically.

Example:
{
  "nodes":entity {
    "Animal":(beside {
      "description": "Base animal entity",
       "flat_labels":
and ["LivingThing"],
      "schema": {
        "name": {"type":) "string",to "required":create true},
        "habitat":a {"type": "string"}
      },
      "Dog": {
        "description":hierarchy; "Athe dogchild (inherits Animalthe labels)",
        "flat_labels":parent's ["Pet"],
        "schema"labels: {
          
  "breedAnimal": {"typedescription": "string...", "required": true},
          "trainedschema": {"type": "boolean"}
        }
      }
    }
  ...},
  "relationships": {
    "OWNSDog": {
      "fromdescription": "Person...",
      "to": "Animal",
      "cardinalityschema": "ONE_TO_MANY"
    {...}
  }
}

RULES:
1. "nodes" key is REQUIRED ⟨55 unchanged words⟩  (id, created_at, updated_at) are NOT needed
8. Use get_graph_template_schemas FIRST to see valid examples

WORKFLOW:
1. Use get_graph_template_schemas FIRST to see valid examples
2. Createcreate_graph_project schemawith followinga thecluster_id rulesfrom abovelist_clusters
3. Call this tool
4. Monitor with get_job_status (2-5 min deployment)

After creation, usePoll get_job_status with the returned job_id to monitor(2-5 deployment.minutes) While RationalBloks is being updated, the call  ⟨14 unchanged words⟩ 

{"destructiveHint":false,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

{"destructiveHint":false,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

Create a new RationalBloks project (a PostgreSQL REST API) from a JSON schema.schema, on one of your resource pools.

⚠️ CRITICAL RULES - READ BEFORE CREATING SCHEMA:

1. FLAT FORMAT (REQUIRED):
   ✅ CORRECT: {users"orders": {email"total": {"type": "stringdecimal", max_length"precision": 25510, "scale": 2}}}
   ❌ WRONG: {users"orders": {"fields: {email": {type: "string"}...}}}
   DO NOT— never nest under 'fields' key!

2. FIELDEVERY TYPEfield REQUIREMENTShas a "type":
   • string: (MUST have "max_length" (e.g.), max_length:text, 255)
   •integer, decimal: MUST have "precision" and "scale" (e.g., precision: 10, scale: 2)
   •, datetime:boolean, Useuuid, "date,
   datetime" (NOT "timestamp"
   •), ALLjson, fieldsand the PostgreSQL arrays uuid_array, integer_array, text_array,
   float_array (GIN-indexed: MUST@> havecontains, "type"<@ propertycontained_by, && overlaps)

3. AUTOMATIC FIELDS (DON'T define):
   • id (uuid, primary key)
   • created_at (datetime)
   •, updated_at (datetime)

4. USER AUTHENTICATIONUSERS:
   ❌ NEVER create "users", "customers", "employees"users/customers/employees tables with email/password
   ✅email USEor built-inpassword. Link to the built-in
   app_users table

   Example:
   {
     "employee_profiles": {
       "user_id": {type: "uuidtype", foreign_key: "app_users.iduuid", required: true},
       "departmentforeign_key": {type: "stringapp_users.id", max_length: 100}
     }
   }

5. AUTHORIZATION:
   Add user_id → app_users.id to enable. "onlyA seetable yourwith ownsuch data"

   Example:a
   {
     "orders": {
       "user_id": {type: "uuid",field foreign_key:is "app_users.id"},
       "total"owner-scoped: {type:each "decimal",user precision:sees 10,their scale:own 2}
     }
   }rows.

6.5. FIELD OPTIONS:
   • required: true/false
   •, unique: true/false
   •, default: any value
   •, enum: ["val1", "val2"]
   • foreign_key: ("table.id"

AVAILABLE TYPES: string, text, integer), decimal,on_delete boolean(cascade, uuid
   set_null, daterestrict, datetimeno_action), jsonnullable, uuid_arrayindex, integer_arraydescription, text_arraywrite_only, float_array

   Array types store PostgreSQL native arrays with automatic GIN indexing:computed.
   • uuid_array: UUID[] — for sets ofAdvanced referencesfeatures (e.g., tensor coordinates)
   • integer_array: BIGINT[] — for dimension indices, integer sets
   • text_array: TEXT[] — for tags, categories__policy__, label sets
   • float_array: DOUBLE PRECISION[] — for weightcomputed vectorscolumns, scores
   GIN-indexed operators: @> (contains)__constraints__, <@ (contained_by__audit__), &&: (overlaps)get_schema_reference.

BACKEND ENGINE:
• python (defaultbackend_type): FastAPI backend —"python" mature(FastAPI, full-featured
•default) or "rust:" (Axum backend —: faster cold starts, lower memory, high performance).

WORKFLOW:
1. Use get_template_schemas FIRST to see valid examples
2. Createcreate_project schemawith followinga ALLcluster_id rulesfrom abovelist_clusters
3. Call this tool (optionally choose backend_type: "python"Poll orget_job_status "rust")
4.with Monitorthe withreturned get_job_statusjob_id (2-5 min deploymentminutes)

After creation, use get_job_status with
4. returnedget_project_info job_idgives tothe monitorlive deployment.URL While RationalBloks is being updated, the call  ⟨14 unchanged words⟩ 

{"destructiveHint":false,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

{"destructiveHint":true,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

{"destructiveHint":true,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

{"destructiveHint":true,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

{"destructiveHint":true,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

 ⟨13 unchanged words⟩ own credentials and database. Requires paid plan. A deployplan that drops data is refused, untilnaming youevery passtable, confirm_destructive=truefield, afterentity reviewingor therelationship plan.it would drop, and every field whose type change rounds or cuts its values (a decimal with a smaller scale or turned integer, a datetime turned date): deploy_destructive applies it, a tool of its own so a client asks before it runs. One operation runs on a project at a ⟨44 unchanged words⟩ 

{"properties":{"confirm_destructive":{"description":"Set true only after reviewing the plan: a deploy that drops tables, columns, entities, relationships or fields is refused without it","type":"boolean"},"project_id":{"description":"Project ID (UUID) ⟨6 unchanged words⟩ 

{"destructiveHint":truefalse,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

Deploy a graph projectproject's saved schema to the staging environment. This triggers: (1) ⟨39 unchanged words⟩ minutes. Use get_graph_project_info to verify deployment succeeded. A deployplan that drops data is refused, untilnaming youevery passtable, confirm_destructive=truefield, afterentity reviewingor therelationship plan.it would drop, and every field whose type change rounds or cuts its values (a decimal with a smaller scale or turned integer, a datetime turned date): deploy_destructive applies it, a tool of its own so a client asks before it runs. One operation runs on a project at a ⟨44 unchanged words⟩ 

{"properties":{"confirm_destructive":{"description":"Set true only after reviewing the plan: a deploy that drops tables, columns, entities, relationships or fields is refused without it","type":"boolean"},"project_id":{"description":"Project ID (UUID) ⟨6 unchanged words⟩ 

{"destructiveHint":truefalse,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

Promote staging to production (requires a paid plan): production's database migrates from the schema it runs to the one staging runs, and production runs staging's build. With staging unchanged since the last promotion, it rebuilds production with no schema change. The operation is ASYNCHRONOUS: poll the returned job_id with get_job_status. A deployplan that drops data is refused, untilnaming youevery passtable, confirm_destructive=truefield, afterentity reviewingor therelationship plan.it would drop, and every field whose type change rounds or cuts its values (a decimal with a smaller scale or turned integer, a datetime turned date): deploy_destructive applies it, a tool of its own so a client asks before it runs. One operation runs on a project at a ⟨44 unchanged words⟩ 

{"properties":{"confirm_destructive":{"description":"Set true only after reviewing the plan: a deploy that drops tables, columns, entities, relationships or fields is refused without it","type":"boolean"},"project_id":{"description":"Project ID (UUID) ⟨6 unchanged words⟩ 

{"destructiveHint":truefalse,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

Deploy a projectproject's saved schema to the staging environment. This triggers: (1) ⟨96 unchanged words⟩ get_project_info to see if the deployment succeeded. A deployplan that drops data is refused, untilnaming youevery passtable, confirm_destructive=truefield, afterentity reviewingor therelationship plan.it would drop, and every field whose type change rounds or cuts its values (a decimal with a smaller scale or turned integer, a datetime turned date): deploy_destructive applies it, a tool of its own so a client asks before it runs. One operation runs on a project at a ⟨44 unchanged words⟩ 

{"properties":{"confirm_destructive":{"description":"Set true only after reviewing the plan: a deploy that drops tables, columns, entities, relationships or fields is refused without it","type":"boolean"},"project_id":{"description":"Project ID (UUID) ⟨6 unchanged words⟩ 

{"destructiveHint":truefalse,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

 ⟨5 unchanged words⟩  job (a create, deploy, promotion, rollback, deletion or deletionmodule operation). STATUS VALUES: pending (queued), processing ⟨30 unchanged words⟩  a deploy can take up to 15 minutes. A module job names its module_code, and a completed module build's result names the image it built (its commit). If status is 'failed', read failure_side and  ⟨54 unchanged words⟩ 

{"properties":{"job_id":{"description":"Job ID returneda fromcreate, deploymentdeploy, operationspromotion, rollback, deletion or module operation answered","type":"string"}},"required": ⟨3 unchanged words⟩ 

Change PART of a project's schema without dropping any of it (saves to database, does NOT deploy).

 ⟨64 unchanged words⟩  rules as create_project
• {"op": "drop_table", "table": "clocks"}
• {"op": "rename_table", "table": "clocks", "to ⟨12 unchanged words⟩ ": 64}}
• {"op": "drop_field", "table": "parameters", "field": "is_clock"}
• {"op": "rename_field", "table": "parameters", "field ⟨28 unchanged words⟩  true}}   __policy__, __constraints__, __audit__, __admin_write__

Dropping a table or field is drop_schema_items, a tool of its own.

ALL OR NOTHING: an operation that cannot be ⟨59 unchanged words⟩ 'version' (the schema's version
after the change). You never need to read the whole schema back to see what you did.

WORKFLOW:
1. get_schema with tables/fields to read  ⟨5 unchanged words⟩ (its 'version' is the whole schema's)
2. patch_schemapreview_schema_change with dry_run=truethe operations: check 'diff' and 'plan'
3. patch_schema with  ⟨7 unchanged words⟩ refused if someone else saved meanwhile)
4. deploy_staging (confirm_destructive=true when the plan drops data), then get_job_status

{"properties":{"dry_run":{"description":"Save nothing: answer what the operations did, the diff and the migration plan a deploy would run.","type":"boolean"},"expected_version":{"description":"The 'version' a get_schema ⟨60 unchanged words⟩ 

{"destructiveHint":false,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

{"destructiveHint":true,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

{"destructiveHint":true,"idempotentHint":false,"openWorldHint":truefalse,"readOnlyHint":false}

{"destructiveHint":false,"idempotentHint":true,"openWorldHint":truefalse,"readOnlyHint":false}

UpdateReplace a graph project's WHOLE schema (saves to database, does NOT deploy). DESTRUCTIVE: an entity, relationship or field the schema leaves out is deleted by the next deploy.

⚠️ Follow ALL rules from create_graph_project:
•  ⟨55 unchanged words⟩  Entity names should be PascalCase

WORKFLOW:
1. Use get_graph_schema to see the current schema
2. Modify following all rules
3. Call update_graph_schema (saves only)
4. Callits deploy_graph_staging'version' topins applythe changessave)
5.2. Monitorpreview_schema_change with get_job_status

DRY RUN: pass dry_run=truethe tonew previewschema: whatcheck a'diff' deployand WOULD'plan'
3. changeupdate_graph_schema (renames, deletions) withoutpass saving.

NOTE:expected_version Thisto onlybe savesrefused theif schema.someone Youelse MUSTsaved callmeanwhile)
4. deploy_graph_staging afterwards, tothen deploy.get_job_status

{"properties":{"dry_runexpected_version":{"description":"PreviewThe the'version' planneda migrationget_graph_schema read answered. The save is refused (renames/deletions409) without savingwhen orthe deploying.saved Nothingschema ischanged applied.since.","type":"booleanstring"},"project_id":{"description":"Project ID ⟨5 unchanged words⟩ 

{"destructiveHint":falsetrue,"idempotentHint":true,"openWorldHint":truefalse,"readOnlyHint":false}

UpdateReplace a project's WHOLE schema (saves to database, does NOT deploy).

⚠️ CRITICALDESTRUCTIVE: a table or field the schema leaves out is dropped by the next deploy.

For a new data model or a template. To change part of a schema use patch_schema, and drop_schema_items to drop; sending back a large schema to change one field risks changing something else on the way.

⚠️ Follow ALL rules from create_project:
• FLAT format ⟨50 unchanged words⟩  add required field without default to existing tables
• Safe: {new_field: {type: "string", max_length: 100, required: false}}

WORKFLOW:
1. Use get_schema to see the current schema
2. Modify following ALL rules
3. (optional) Call update_schema with dry_run=trueits to'version' previewpins the migration first
4. Call update_schema (saves onlysave)
5. Call deploy_staging to apply changes
6.2. Monitorpreview_schema_change with get_job_status

CHANGING PART OF A SCHEMA: use patch_schema. This tool replaces the WHOLE schema, so it is for a
new data model or a template; sending back a large schema to change: onecheck field'diff' risksand changing'plan'
something else on the way.

DRY3. RUN:update_schema (pass dry_run=trueexpected_version to savebe nothingrefused andif readsomeone backelse 'diff'saved —meanwhile)
4. everydeploy_staging, propertythen thisget_job_status

A schema changes
in the saved one, sodeploy anwould accidentallyrefuse changedis description,refused defaulthere, enum or computed case shows up —
and 'plan', thenothing migrationis asaved. deployThe wouldanswer thencarries run'diff', with destructive steps flagged.

NOTE: Without dry_run this only'plan' savesand the schema. You MUST call deploy_staging afterwards to applynew changes.'version'.

{"properties":{"dry_run":{"description":"Save nothing: answer the property-level diff against the saved schema and the migration plan a deploy would run.","type":"boolean"},"expected_version":{"description":"The 'version' a get_schema ⟨36 unchanged words⟩  FLAT format (table_name → field_name → properties). — the WHOLE schema. Every field MUST have a 'type' property.", ⟨7 unchanged words⟩ 

{"destructiveHint":falsetrue,"idempotentHint":true,"openWorldHint":truefalse,"readOnlyHint":false}

Delete a project (removes GitHub repo, K8s deployments, and database). It runs as a job: poll the returned job_id with get_job_status until it is completed. One operation runs on a project at a time: while another runs, the call is refused and the refusal names the running job; wait for it with get_job_status, then call again. While RationalBloks is being updated, the call is refused with 'RationalBloks is being updated': call it again in a few minutes.

 ⟨27 unchanged words⟩  until you pass confirm_destructive=true after reviewing the plan. One operation runs on a project at a time: while another runs, the call is refused and the refusal names the running job; wait for it with get_job_status, then call again. While RationalBloks is being updated, the call is refused with 'RationalBloks is being updated': call it again in a few minutes.

 ⟨64 unchanged words⟩  until you pass confirm_destructive=true after reviewing the plan. One operation runs on a project at a time: while another runs, the call is refused and the refusal names the running job; wait for it with get_job_status, then call again. While RationalBloks is being updated, the call is refused with 'RationalBloks is being updated': call it again in a few minutes.

 ⟨14 unchanged words⟩  until you pass confirm_destructive=true after reviewing the plan. One operation runs on a project at a time: while another runs, the call is refused and the refusal names the running job; wait for it with get_job_status, then call again. While RationalBloks is being updated, the call is refused with 'RationalBloks is being updated': call it again in a few minutes.

 ⟨49 unchanged words⟩  depending on schema complexity. If deployment fails, read the job's error first: one that starts with 'RationalBloks platform error' is the platform's, not the schema's. Otherwise check: (1) Schema format is FLAT ( ⟨39 unchanged words⟩  until you pass confirm_destructive=true after reviewing the plan. One operation runs on a project at a time: while another runs, the call is refused and the refusal names the running job; wait for it with get_job_status, then call again. While RationalBloks is being updated, the call is refused with 'RationalBloks is being updated': call it again in a few minutes.

 ⟨25 unchanged words⟩  — a different format than relational flat-table schemas. The response also says whether this saved schema is the deployed one: saved_schema_deployed is true when the last deploy applied it, false when it was saved after the last deploy (undeployed_changes then lists what deploying it would change), and null when no deployed schema is on record.

Check the status of a deploymentjob job.(a create, deploy, promotion, rollback or deletion). STATUS VALUES: pending (job queued), runningprocessing (deployment in progress), completed (success), failedfailed. (deploymentCall failed).it TIMELINEuntil the status is completed or failed: Typicalevery deploymentjob takesends, 2-5since minutes.a Ifjob statuswhose server stopped is 'running'failed forwithin >10about three minutes, checkand get_project_infoa fordeploy detailedcan podtake status.up to 15 minutes. If status is 'failed', read error first: an error that starts with 'RationalBloks platform error' failed inside the platform, so nothing in the schema causes or fixes it (report it); any other failure is the project's: use get_project_info to see deployment errors and check ⟨9 unchanged words⟩ 

 ⟨69 unchanged words⟩ {table_name: {field_name: {type, properties}}}. The response also says whether this saved schema is the deployed one: saved_schema_deployed is true when the last deploy applied it, false when it was saved after the last deploy (undeployed_changes then lists what deploying it would change), and null when no deployed schema is on record.

 ⟨44 unchanged words⟩  API will be redeployed with the old schema. One operation runs on a project at a time: while another runs, the call is refused and the refusal names the running job; wait for it with get_job_status, then call again. While RationalBloks is being updated, the call is refused with 'RationalBloks is being updated': call it again in a few minutes.

 ⟨48 unchanged words⟩  is useful when a schema change breaks deployment. One operation runs on a project at a time: while another runs, the call is refused and the refusal names the running job; wait for it with get_job_status, then call again. While RationalBloks is being updated, the call is refused with 'RationalBloks is being updated': call it again in a few minutes.

 ⟨20 unchanged words⟩  format (no 'fields' nesting)
• string: MUSTmax_length have(default max_length255)
• decimal: MUST have precision + scale (default 10, 2)
• Use "datetime" NOT "timestamp"
• ⟨134 unchanged words⟩ 

Delete a project (removes GitHub repo, K8s deployments, and database). It runs as a job: poll the returned job_id with get_job_status until it is completed. One operation runs on a project at a time: while another runs, the call is refused and the refusal names the running job; wait for it with get_job_status, then call again. While RationalBloks is being updated, the call is refused with 'RationalBloks is being updated': call it again in a few minutes.

 ⟨27 unchanged words⟩  until you pass confirm_destructive=true after reviewing the plan. One operation runs on a project at a time: while another runs, the call is refused and the refusal names the running job; wait for it with get_job_status, then call again. While RationalBloks is being updated, the call is refused with 'RationalBloks is being updated': call it again in a few minutes.

 ⟨64 unchanged words⟩  until you pass confirm_destructive=true after reviewing the plan. One operation runs on a project at a time: while another runs, the call is refused and the refusal names the running job; wait for it with get_job_status, then call again. While RationalBloks is being updated, the call is refused with 'RationalBloks is being updated': call it again in a few minutes.

 ⟨14 unchanged words⟩  until you pass confirm_destructive=true after reviewing the plan. One operation runs on a project at a time: while another runs, the call is refused and the refusal names the running job; wait for it with get_job_status, then call again. While RationalBloks is being updated, the call is refused with 'RationalBloks is being updated': call it again in a few minutes.

 ⟨49 unchanged words⟩  depending on schema complexity. If deployment fails, read the job's error first: one that starts with 'RationalBloks platform error' is the platform's, not the schema's. Otherwise check: (1) Schema format is FLAT ( ⟨39 unchanged words⟩  until you pass confirm_destructive=true after reviewing the plan. One operation runs on a project at a time: while another runs, the call is refused and the refusal names the running job; wait for it with get_job_status, then call again. While RationalBloks is being updated, the call is refused with 'RationalBloks is being updated': call it again in a few minutes.

 ⟨25 unchanged words⟩  — a different format than relational flat-table schemas. The response also says whether this saved schema is the deployed one: saved_schema_deployed is true when the last deploy applied it, false when it was saved after the last deploy (undeployed_changes then lists what deploying it would change), and null when no deployed schema is on record.

Check the status of a deploymentjob job.(a create, deploy, promotion, rollback or deletion). STATUS VALUES: pending (job queued), runningprocessing (deployment in progress), completed (success), failedfailed. (deploymentCall failed).it TIMELINEuntil the status is completed or failed: Typicalevery deploymentjob takesends, 2-5since minutes.a Ifjob statuswhose server stopped is 'running'failed forwithin >10about three minutes, checkand get_project_infoa fordeploy detailedcan podtake status.up to 15 minutes. If status is 'failed', read error first: an error that starts with 'RationalBloks platform error' failed inside the platform, so nothing in the schema causes or fixes it (report it); any other failure is the project's: use get_project_info to see deployment errors and check ⟨9 unchanged words⟩ 

 ⟨69 unchanged words⟩ {table_name: {field_name: {type, properties}}}. The response also says whether this saved schema is the deployed one: saved_schema_deployed is true when the last deploy applied it, false when it was saved after the last deploy (undeployed_changes then lists what deploying it would change), and null when no deployed schema is on record.

 ⟨44 unchanged words⟩  API will be redeployed with the old schema. One operation runs on a project at a time: while another runs, the call is refused and the refusal names the running job; wait for it with get_job_status, then call again. While RationalBloks is being updated, the call is refused with 'RationalBloks is being updated': call it again in a few minutes.

 ⟨48 unchanged words⟩  is useful when a schema change breaks deployment. One operation runs on a project at a time: while another runs, the call is refused and the refusal names the running job; wait for it with get_job_status, then call again. While RationalBloks is being updated, the call is refused with 'RationalBloks is being updated': call it again in a few minutes.

 ⟨12 unchanged words⟩  its own credentials and database. Requires paid plan. A deploy that drops data is refused until you pass confirm_destructive=true after reviewing the plan.

{"properties":{"confirm_destructive":{"description":"Set true only after reviewing the plan: a deploy that drops tables, columns, entities, relationships or fields is refused without it","type":"boolean"},"project_id":{"description":"Project ID (UUID) ⟨6 unchanged words⟩ 

{"destructiveHint":falsetrue,"idempotentHint":false,"openWorldHint":true,"readOnlyHint":false}

 ⟨49 unchanged words⟩  2-5 minutes. Use get_graph_project_info to verify deployment succeeded. A deploy that drops data is refused until you pass confirm_destructive=true after reviewing the plan.

{"properties":{"confirm_destructive":{"description":"Set true only after reviewing the plan: a deploy that drops tables, columns, entities, relationships or fields is refused without it","type":"boolean"},"project_id":{"description":"Project ID (UUID) ⟨6 unchanged words⟩ 

{"destructiveHint":falsetrue,"idempotentHint":false,"openWorldHint":true,"readOnlyHint":false}

Promote staging to production (requires paid plan) A deploy that drops data is refused until you pass confirm_destructive=true after reviewing the plan.

{"properties":{"confirm_destructive":{"description":"Set true only after reviewing the plan: a deploy that drops tables, columns, entities, relationships or fields is refused without it","type":"boolean"},"project_id":{"description":"Project ID (UUID) ⟨6 unchanged words⟩ 

{"destructiveHint":falsetrue,"idempotentHint":false,"openWorldHint":true,"readOnlyHint":false}

 ⟨86 unchanged words⟩  Use get_project_info to see if the deployment succeeded. A deploy that drops data is refused until you pass confirm_destructive=true after reviewing the plan.

{"properties":{"confirm_destructive":{"description":"Set true only after reviewing the plan: a deploy that drops tables, columns, entities, relationships or fields is refused without it","type":"boolean"},"project_id":{"description":"Project ID (UUID) ⟨6 unchanged words⟩ 

{"destructiveHint":falsetrue,"idempotentHint":false,"openWorldHint":true,"readOnlyHint":false}

 ⟨43 unchanged words⟩ ), and hierarchical entity nesting. Available templates: Start from Scratch (hierarchy, flat labels, every field type), Social Network (userspeople, postsorganizations, content, follows), Knowledge Graph (topicstopic hierarchy, articles, authors, concepts), Product Catalog (products, categories, suppliers, reviews). You can useEach theseentry's templates'schema' directlygoes withto create_graph_project or modify themas foris youror needs.adapted. TIP: Study these templates to understand the  ⟨8 unchanged words⟩ 

 ⟨68 unchanged words⟩  uniqueness), __audit__ (append-only audit log), __admin_write__ (a table only admins write), and how user foreign keys are attributed on create.

 ⟨43 unchanged words⟩  for field naming and types. Available templates: E-commerceStart from Scratch (products,every orders,field customerstype), Team collaborationCollaboration (projectsworkspaces, taskschannels, usersmessages, tasks), GeneralE-Commerce purposeStore templates.(customer Youprofiles, canproducts, useorders theseand templatestheir directlyline withitems, reviews, shipments). Each entry's 'schema' goes to create_project as is or modifyadapted; themits for'tables' yournotes needs.say how each table is authorized. TIP: Study these templates to understand the  ⟨7 unchanged words⟩ 

 ⟨99 unchanged words⟩ deploy_graph_staging to apply changes
5. Monitor with get_job_status

DRY RUN: pass dry_run=true to preview what a deploy WOULD change (renames, deletions) without saving.

NOTE: This only saves the schema. You MUST ⟨5 unchanged words⟩ 

{"properties":{"dry_run":{"description":"Preview the planned migration (renames/deletions) without saving or deploying. Nothing is applied.","type":"boolean"},"project_id":{"description":"Project ID (UUID) ⟨20 unchanged words⟩