settlement mode updates

README.md

@@ -439,7 +439,6 @@ OpenGradient supports multiple settlement modes through the x402 payment protoco
 
 - **PRIVATE**: Payment only, no input/output data on-chain (maximum privacy)
 - **BATCH_HASHED**: Aggregates inferences into a Merkle tree with input/output hashes and signatures (most cost-efficient, default)
-- **INDIVIDUAL_FULL**: Records input, output, timestamp, and verification on-chain (maximum auditability)
 
 Specify settlement mode in your requests:
 ```python

docs/CLAUDE_SDK_USERS.md

@@ -282,9 +282,8 @@ og.InferenceMode.TEE        # Trusted Execution Environment
 og.InferenceMode.ZKML       # Zero-knowledge proof
 
 # x402 Payment Settlement Modes (for LLM calls)
-og.x402SettlementMode.PRIVATE           # Input/output hashes only (most private)
-og.x402SettlementMode.BATCH_HASHED     # Batch hashes (most cost-efficient, default)
-og.x402SettlementMode.INDIVIDUAL_FULL  # Full data and metadata on-chain
+og.x402SettlementMode.PRIVATE       # Payment only; inference data stays off-chain
+og.x402SettlementMode.BATCH_HASHED  # Batch hashes (most cost-efficient, default)
 
 # Workflow data types
 og.CandleType.OPEN, .HIGH, .LOW, .CLOSE, .VOLUME

docs/opengradient/client/llm.md

@@ -99,7 +99,6 @@ Perform inference on an LLM model using chat via TEE.
 * **`x402_settlement_mode (x402SettlementMode, optional)`**: Settlement mode for x402 payments.
         - PRIVATE: Payment only, no input/output data on-chain (most privacy-preserving).
         - BATCH_HASHED: Aggregates inferences into a Merkle tree with input/output hashes and signatures (default, most cost-efficient).
-        - INDIVIDUAL_FULL: Records input, output, timestamp, and verification on-chain (maximum auditability).
         Defaults to BATCH_HASHED.
 * **`stream (bool, optional)`**: Whether to stream the response. Default is False.
 
@@ -115,9 +114,8 @@ Union[TextGenerationOutput, AsyncGenerator[StreamChunk, None]]:
 * **`data_settlement_transaction_hash`**: Blockchain transaction hash for
         the data settlement transaction. ``None`` when the provider
         does not return data settlement metadata.
-* **`data_settlement_blob_id`**: Walrus blob ID for individual data
-        settlement. ``None`` for private/batch settlement or when the
-        provider does not return it.
+* **`data_settlement_blob_id`**: Walrus blob ID for data settlement.
+        ``None`` when the provider does not return it.
 * **`finish_reason`**: Reason the model stopped generating
         (e.g. ``"stop"``, ``"tool_call"``, ``"error"``).
         Only populated for chat requests.
@@ -179,7 +177,6 @@ Perform inference on an LLM model using completions via TEE.
 * **`x402_settlement_mode (x402SettlementMode, optional)`**: Settlement mode for x402 payments.
         - PRIVATE: Payment only, no input/output data on-chain (most privacy-preserving).
         - BATCH_HASHED: Aggregates inferences into a Merkle tree with input/output hashes and signatures (default, most cost-efficient).
-        - INDIVIDUAL_FULL: Records input, output, timestamp, and verification on-chain (maximum auditability).
         Defaults to BATCH_HASHED.
 
 **Returns**
@@ -194,9 +191,8 @@ TextGenerationOutput: Generated text results including:
 * **`data_settlement_transaction_hash`**: Blockchain transaction hash for
         the data settlement transaction. ``None`` when the provider
         does not return data settlement metadata.
-* **`data_settlement_blob_id`**: Walrus blob ID for individual data
-        settlement. ``None`` for private/batch settlement or when the
-        provider does not return it.
+* **`data_settlement_blob_id`**: Walrus blob ID for data settlement.
+        ``None`` when the provider does not return it.
 * **`finish_reason`**: Reason the model stopped generating
         (e.g. ``"stop"``, ``"tool_call"``, ``"error"``).
         Only populated for chat requests.
@@ -269,4 +265,4 @@ Resolve the current TEE or a specific active registry TEE.
 
 This is primarily for backend relays that need SDK-managed TEE routing,
 TLS pinning, and x402 clients without using the chat/completion helpers
-directly, for example when forwarding OHTTP ciphertext.
+directly, for example when forwarding OHTTP ciphertext.

docs/opengradient/client/twins.md

@@ -58,9 +58,8 @@ TextGenerationOutput: Generated text results including chat_output and finish_re
 * **`data_settlement_transaction_hash`**: Blockchain transaction hash for
         the data settlement transaction. ``None`` when the provider
         does not return data settlement metadata.
-* **`data_settlement_blob_id`**: Walrus blob ID for individual data
-        settlement. ``None`` for private/batch settlement or when the
-        provider does not return it.
+* **`data_settlement_blob_id`**: Walrus blob ID for data settlement.
+        ``None`` when the provider does not return it.
 * **`finish_reason`**: Reason the model stopped generating
         (e.g. ``"stop"``, ``"tool_call"``, ``"error"``).
         Only populated for chat requests.

docs/opengradient/types.md

@@ -352,7 +352,7 @@ usage information.
 * **`tee_payment_address`**: Payment address registered for the TEE (final chunk only)
 * **`data_settlement_transaction_hash`**: Transaction hash for the data settlement
         transaction, present on the final chunk when available.
-* **`data_settlement_blob_id`**: Walrus blob ID for individual data settlement,
+* **`data_settlement_blob_id`**: Walrus blob ID for data settlement,
         present on the final chunk when available.
 * **`images`**: Generated images returned by image-output models, present on the
         final chunk when available. Each entry is a ``data:`` URI.
@@ -410,7 +410,7 @@ StreamChunk instance
 * **`tee_payment_address`**: Payment address registered for the TEE (final chunk only)
 * **`data_settlement_transaction_hash`**: Transaction hash for the data settlement
         transaction, present on the final chunk when available.
-* **`data_settlement_blob_id`**: Walrus blob ID for individual data settlement,
+* **`data_settlement_blob_id`**: Walrus blob ID for data settlement,
         present on the final chunk when available.
 * **`images`**: Generated images returned by image-output models, present on the
         final chunk when available. Each entry is a ``data:`` URI.
@@ -557,9 +557,8 @@ Trust model:
 * **`data_settlement_transaction_hash`**: Blockchain transaction hash for
         the data settlement transaction. ``None`` when the provider
         does not return data settlement metadata.
-* **`data_settlement_blob_id`**: Walrus blob ID for individual data
-        settlement. ``None`` for private/batch settlement or when the
-        provider does not return it.
+* **`data_settlement_blob_id`**: Walrus blob ID for data settlement.
+        ``None`` when the provider does not return it.
 * **`finish_reason`**: Reason the model stopped generating
         (e.g. ``"stop"``, ``"tool_call"``, ``"error"``).
         Only populated for chat requests.
@@ -633,9 +632,4 @@ privacy, and transaction costs.
         Aggregates multiple inferences into a single settlement transaction
         using a Merkle tree containing input hashes, output hashes, and signatures.
         Most cost-efficient for high-volume applications.
-        CLI usage: --settlement-mode batch-hashed
-* **`INDIVIDUAL_FULL`**: Individual settlement with full metadata.
-        Records input data, output data, timestamp, and verification on-chain.
-        Provides maximum transparency and auditability.
-        Higher gas costs due to larger data storage.
-        CLI usage: --settlement-mode individual-full
+        CLI usage: --settlement-mode batch-hashed

examples/langchain_react_agent.py

@@ -26,7 +26,7 @@
     private_key=private_key,
     model_cid=og.TEE_LLM.GPT_4_1_2025_04_14,
     max_tokens=300,
-    x402_settlement_mode=og.x402SettlementMode.INDIVIDUAL_FULL,
+    x402_settlement_mode=og.x402SettlementMode.BATCH_HASHED,
 )
 
 

examples/llm_chat.py

@@ -16,12 +16,12 @@ async def main():
         {"role": "user", "content": "What model are you?"},
     ]
 
-    # Run inference with full public settlement
+    # Run inference with batched settlement
     result = await llm.chat(
         model=og.TEE_LLM.CLAUDE_OPUS_4_8,
         messages=messages,
         max_tokens=300,
-        x402_settlement_mode=og.x402SettlementMode.INDIVIDUAL_FULL,
+        x402_settlement_mode=og.x402SettlementMode.BATCH_HASHED,
     )
     print(result.chat_output["content"])
 

examples/llm_chat_streaming.py

@@ -14,7 +14,7 @@ async def main():
         {"role": "user", "content": "What makes it good for beginners?"},
     ]
 
-    settlement_mode = og.x402SettlementMode.INDIVIDUAL_FULL
+    settlement_mode = og.x402SettlementMode.BATCH_HASHED
     stream = await llm.chat(
         model=og.TEE_LLM.GPT_4_1_2025_04_14,
         messages=messages,
@@ -27,11 +27,10 @@ async def main():
         if chunk.choices[0].delta.content:
             print(chunk.choices[0].delta.content, end="", flush=True)
 
-        if settlement_mode == og.x402SettlementMode.INDIVIDUAL_FULL:
-            if chunk.data_settlement_blob_id:
-                print("\n Data Settlement Blob ID: ", chunk.data_settlement_blob_id)
-            if chunk.data_settlement_transaction_hash:
-                print("\n Data Settlement Transaction Hash: ", chunk.data_settlement_transaction_hash)
+        if chunk.data_settlement_blob_id:
+            print("\n Data Settlement Blob ID: ", chunk.data_settlement_blob_id)
+        if chunk.data_settlement_transaction_hash:
+            print("\n Data Settlement Transaction Hash: ", chunk.data_settlement_transaction_hash)
 
 
 asyncio.run(main())

integrationtest/llm/test_llm_chat.py

@@ -135,7 +135,7 @@ async def test_chat(llm_client):
         model=og.TEE_LLM.GEMINI_2_5_FLASH,
         messages=messages,
         max_tokens=50,
-        x402_settlement_mode=og.x402SettlementMode.INDIVIDUAL_FULL,
+        x402_settlement_mode=og.x402SettlementMode.BATCH_HASHED,
     )
 
     assert result is not None
@@ -152,7 +152,7 @@ async def test_chat_streaming(llm_client):
         model=og.TEE_LLM.GEMINI_2_5_FLASH,
         messages=messages,
         max_tokens=50,
-        x402_settlement_mode=og.x402SettlementMode.INDIVIDUAL_FULL,
+        x402_settlement_mode=og.x402SettlementMode.BATCH_HASHED,
         stream=True,
     )
 

src/opengradient/cli.py

@@ -73,7 +73,6 @@ def convert(self, value, param, ctx):
 x402SettlementModes = {
     "batch-hashed": x402SettlementMode.BATCH_HASHED,
     "private": x402SettlementMode.PRIVATE,
-    "individual-full": x402SettlementMode.INDIVIDUAL_FULL,
 }
 
 
@@ -384,7 +383,7 @@ def infer(ctx, model_cid: str, inference_mode: str, input_data, input_file: Path
     "x402_settlement_mode",
     type=click.Choice(x402SettlementModes.keys()),
     default="batch-hashed",
-    help="Settlement mode for x402 payments: private (payment only), batch-hashed (default), individual-full (full data with verification)",
+    help="Settlement mode for x402 payments: private (payment only), batch-hashed (default)",
 )
 @click.pass_context
 def completion(
@@ -518,7 +517,7 @@ def print_llm_completion_result(model_cid, tx_hash, llm_output, is_vanilla=True,
     "--x402-settlement-mode",
     type=click.Choice(x402SettlementModes.keys()),
     default="batch-hashed",
-    help="Settlement mode for x402 payments: private (payment only), batch-hashed (default), individual-full (full data with verification)",
+    help="Settlement mode for x402 payments: private (payment only), batch-hashed (default)",
 )
 @click.option("--stream", is_flag=True, default=False, help="Stream the output from the LLM")
 @click.pass_context

src/opengradient/client/llm.py

@@ -339,7 +339,6 @@ async def completion(
             x402_settlement_mode (x402SettlementMode, optional): Settlement mode for x402 payments.
                 - PRIVATE: Payment only, no input/output data on-chain (most privacy-preserving).
                 - BATCH_HASHED: Aggregates inferences into a Merkle tree with input/output hashes and signatures (default, most cost-efficient).
-                - INDIVIDUAL_FULL: Records input, output, timestamp, and verification on-chain (maximum auditability).
                 Defaults to BATCH_HASHED.
 
         Returns:
@@ -427,7 +426,6 @@ async def chat(
             x402_settlement_mode (x402SettlementMode, optional): Settlement mode for x402 payments.
                 - PRIVATE: Payment only, no input/output data on-chain (most privacy-preserving).
                 - BATCH_HASHED: Aggregates inferences into a Merkle tree with input/output hashes and signatures (default, most cost-efficient).
-                - INDIVIDUAL_FULL: Records input, output, timestamp, and verification on-chain (maximum auditability).
                 Defaults to BATCH_HASHED.
             stream (bool, optional): Whether to stream the response. Default is False.
 

src/opengradient/types.py

@@ -31,12 +31,6 @@ class x402SettlementMode(str, Enum):
             Most cost-efficient for high-volume applications.
             CLI usage: --settlement-mode batch-hashed
 
-        INDIVIDUAL_FULL: Individual settlement with full metadata.
-            Records input data, output data, timestamp, and verification on-chain.
-            Provides maximum transparency and auditability.
-            Higher gas costs due to larger data storage.
-            CLI usage: --settlement-mode individual-full
-
     Examples:
         >>> from opengradient import x402SettlementMode
         >>> mode = x402SettlementMode.PRIVATE
@@ -46,7 +40,6 @@ class x402SettlementMode(str, Enum):
 
     PRIVATE = "private"
     BATCH_HASHED = "batch"
-    INDIVIDUAL_FULL = "individual"
 
 
 class CandleOrder(IntEnum):
@@ -242,7 +235,7 @@ class StreamChunk:
         tee_payment_address: Payment address registered for the TEE (final chunk only)
         data_settlement_transaction_hash: Transaction hash for the data settlement
             transaction, present on the final chunk when available.
-        data_settlement_blob_id: Walrus blob ID for individual data settlement,
+        data_settlement_blob_id: Walrus blob ID for data settlement,
             present on the final chunk when available.
         images: Generated images returned by image-output models, present on the
             final chunk when available. Each entry is a ``data:`` URI.
@@ -430,9 +423,8 @@ class TextGenerationOutput:
         data_settlement_transaction_hash: Blockchain transaction hash for
             the data settlement transaction. ``None`` when the provider
             does not return data settlement metadata.
-        data_settlement_blob_id: Walrus blob ID for individual data
-            settlement. ``None`` for private/batch settlement or when the
-            provider does not return it.
+        data_settlement_blob_id: Walrus blob ID for data settlement.
+            ``None`` when the provider does not return it.
         finish_reason: Reason the model stopped generating
             (e.g. ``"stop"``, ``"tool_call"``, ``"error"``).
             Only populated for chat requests.
@@ -453,7 +445,7 @@ class TextGenerationOutput:
     """Blockchain transaction hash for the data settlement transaction. ``None`` when unavailable."""
 
     data_settlement_blob_id: Optional[str] = None
-    """Walrus blob ID for individual data settlement. ``None`` when unavailable."""
+    """Walrus blob ID for data settlement. ``None`` when unavailable."""
 
     finish_reason: Optional[str] = None
     """Reason the model stopped generating (e.g. ``"stop"``, ``"tool_call"``, ``"error"``). Only populated for chat requests."""

stresstest/llm.py

@@ -17,7 +17,7 @@ async def main(private_key: str):
 
     async def run_prompt(prompt: str):
         messages = [{"role": "user", "content": prompt}]
-        await llm.chat(MODEL, messages=messages, max_tokens=50, x402_settlement_mode=og.x402SettlementMode.INDIVIDUAL_FULL)
+        await llm.chat(MODEL, messages=messages, max_tokens=50, x402_settlement_mode=og.x402SettlementMode.BATCH_HASHED)
 
     latencies, failures = await stress_test_wrapper(run_prompt, num_requests=NUM_REQUESTS)
 

tests/client_test.py

@@ -210,4 +210,7 @@ def test_settlement_modes_values(self):
         """Test settlement mode enum values."""
         assert x402SettlementMode.PRIVATE == "private"
         assert x402SettlementMode.BATCH_HASHED == "batch"
-        assert x402SettlementMode.INDIVIDUAL_FULL == "individual"
+        assert list(x402SettlementMode) == [
+            x402SettlementMode.PRIVATE,
+            x402SettlementMode.BATCH_HASHED,
+        ]

tutorials/01-verifiable-ai-agent.md

@@ -220,9 +220,8 @@ is recorded on-chain:
 
 | Mode | What's Stored | Best For |
 |------|--------------|----------|
-| `PRIVATE` | Hashes of input and output only | **Privacy** -- proves execution happened without revealing content |
+| `PRIVATE` | Payment only; no input/output data | **Privacy** -- keeps inference data off-chain |
 | `BATCH_HASHED` | Batch hash of multiple inferences | **Cost efficiency** -- reduces per-call gas costs (default) |
-| `INDIVIDUAL_FULL` | Full model info, input, output, and metadata | **Transparency** -- complete auditability for compliance |
 
 Choose based on your requirements:
 
@@ -234,13 +233,6 @@ llm_dev = og.agents.langchain_adapter(
     x402_settlement_mode=og.x402SettlementMode.BATCH_HASHED,
 )
 
-# For production financial applications -- full audit trail
-llm_prod = og.agents.langchain_adapter(
-    private_key=os.environ["OG_PRIVATE_KEY"],
-    model_cid=og.TEE_LLM.GPT_4_1_2025_04_14,
-    x402_settlement_mode=og.x402SettlementMode.INDIVIDUAL_FULL,
-)
-
 # For privacy-sensitive applications -- minimal on-chain footprint
 llm_private = og.agents.langchain_adapter(
     private_key=os.environ["OG_PRIVATE_KEY"],
@@ -354,5 +346,5 @@ if __name__ == "__main__":
   models deployed on OpenGradient.
 - **Read workflow results**: Use `og.alphasense.create_read_workflow_tool` to read
   from scheduled on-chain workflows that run models automatically.
-- **Go to production**: Switch settlement mode to `INDIVIDUAL_FULL` and store the
-  payment hashes and transaction hashes for your compliance records.
+- **Go to production**: Store payment hashes and transaction hashes for your
+  compliance records.