GITBOOK-201: Add Streaming option documentation

Author: Artem Grigor

docs/.gitbook/assets/tmp.md

@@ -0,0 +1,49 @@
+
+
+### Streaming Options
+
+Tycho Client supports several options to customize the data stream. These are available as CLI flags, Rust builder
+methods, and Python parameters — refer to each client's documentation for usage details.
+
+| CLI Flag                | Rust Builder               | Python                     | Description                                                         |
+|-------------------------|----------------------------|----------------------------|---------------------------------------------------------------------|
+| `--partial-blocks`      | `.enable_partial_blocks()` | `partial_blocks=True`      | Enable partial block updates (flash blocks) for sub-block latency   |
+| `--no-state`            | `.no_state(true)`          | `include_state=False`      | Only stream components and tokens; omit snapshots and state updates |
+| `--include-tvl`         | `.include_tvl(true)`       | `include_tvl=True`         | Include TVL estimates in messages                                   |
+| `--disable-compression` | `.disable_compression()`   | `disable_compression=True` | Disable zstd compression for WebSocket messages                     |
+| `--no-tls`              | `.no_tls(true)`            | `use_tls=False`            | Use unencrypted transports (http/ws instead of https/wss)           |
+
+{% hint style="info" %}
+**Not recommended:** `--disable-compression` and `--no-tls` are intended for debugging only. Compression (zstd)
+significantly reduces message sizes and improves throughput, while TLS is required to connect to the hosted endpoint.
+Keeping both enabled (the default) is recommended for production use.
+{% endhint %}
+
+#### Partial Blocks
+
+Some chains, such as [Base](https://docs.base.org/building-with-base/differences/flashblocks), support _flash blocks_ —
+pre-confirmation updates that contain parts of a future block before it is finalized. When `--partial-blocks` is
+enabled, Tycho streams these incremental updates as they arrive, giving you sub-block latency. On chains without flash block
+support, enabling this flag has no effect — full blocks are delivered as usual.
+
+{% hint style="warning" %}
+Block hashes in partial block messages are **unstable** — they change between partial updates and will differ from the
+final block hash. Do not use them as persistent identifiers or cache keys. See
+the [Substreams documentation](https://docs.substreams.dev/reference-material/chain-support/flashblocks#developing-for-partial-blocks)
+for details.
+{% endhint %}
+
+#### No State
+
+By default, each message includes full component snapshots (on the first sync) and state deltas (reserves, balances,
+contract storage) on every subsequent block. If you only need to discover which components exist and which tokens they
+involve — for example, to build a pool registry or monitor new deployments — you can disable state with `--no-state`.
+This significantly reduces message sizes and processing overhead since snapshots and per-block state updates are omitted
+entirely.
+
+#### Include TVL
+
+When enabled, each message includes an approximate TVL estimate for every tracked component. This is useful for building
+dashboards or for ranking pools by liquidity. Note that enabling this option increases startup latency: for each
+snapshot request, the client makes additional RPC calls to fetch token prices and compute TVL for all tracked
+components. This overhead scales with the number of components you're tracking.

docs/for-dexs/protocol-integration/3.-testing.md

@@ -29,9 +29,10 @@ You need an EVM **Archive** node to fetch the state from a previous block. If yo
 
 The node also needs to support the [debug\_storageRangeAt](https://www.quicknode.com/docs/ethereum/debug_storageRangeAt) method, which is required for our Token Quality Analysis.
 
-As of February 2026, Erigon is the only major client supporting debug_storageRangeAt. The following API providers support archive nodes with debug\_storageRangeAt:
-- [Chainnodes](https://www.chainnodes.org/)
-- [Chainstack](https://chainstack.com/)
+As of February 2026, Erigon is the only major client supporting debug\_storageRangeAt. The following API providers support archive nodes with debug\_storageRangeAt:
+
+* [Chainnodes](https://www.chainnodes.org/)
+* [Chainstack](https://chainstack.com/)
 
 ### Test Configuration
 
@@ -129,18 +130,18 @@ By default this should be `false` . It should only be `true` temporarily if you
 
 To be able to test execution, you need to provide the executor's runtime bytecode file.
 
-1. Export it using the helper script in [tycho-execution/foundry/scripts/export-runtime-bytecode.js ](https://github.com/propeller-heads/tycho-execution/blob/main/foundry/scripts/export-runtime-bytecode.js) (see the [README](https://github.com/propeller-heads/tycho-execution/blob/main/foundry/scripts/README.md#export-runtime-bytecode) for instructions on how).
+1. Export it using the helper script in [tycho-execution/foundry/scripts/export-runtime-bytecode.js ](https://github.com/propeller-heads/tycho-execution/blob/main/foundry/scripts/export-runtime-bytecode.js)(see the [README](https://github.com/propeller-heads/tycho-execution/blob/main/foundry/scripts/README.md#export-runtime-bytecode) for instructions on how).
 2. Copy `YourExecutor.runtime.json` file to the SDK repository in [`tycho-protocol-sdk/evm/test/executors`](https://github.com/propeller-heads/tycho-protocol-sdk/tree/main/evm/test/executors) .
 3. Import the file in [tycho-protocol-sdk/protocol-testing/src/execution.rs](https://github.com/propeller-heads/tycho-protocol-sdk/blob/3f31f85c157b1f860fcc9604376e2c11b1e8da0c/protocol-testing/src/execution.rs#L51) and add the corresponding entry to the `EXECUTOR_MAPPING` .
 
 {% hint style="danger" %}
-### Block Compatibility Requirements
+#### Block Compatibility Requirements
 
 The `TychoRouter` requires post-Cancun blocks for execution. Testing must use block numbers after the Cancun upgrade.
 {% endhint %}
 
 {% hint style="info" %}
-#### Testing during development
+**Testing during development**
 
 To test your protocol integration during development, update the `tycho-simulation` and `tycho-execution` dependencies in `protocol-testing/Cargo.toml` to point to your working branch/commit or to your local repository.
 

docs/for-solvers/indexer/tycho-client/README.md

@@ -4,10 +4,16 @@ Tycho Client helps you consume data from Tycho Indexer. It's the recommended way
 
 In this guide, you'll learn more about the Tycho Client and the streamed data models.
 
-{% hint style="info" %}
+{% hint style="success" %}
 If you are developing in Rust and is using Tycho to simulate DeFi Protocol's behavior, we recommend checking out our [simulation.md](../../simulation.md "mention") package - this tool extends Tycho Client's data streaming functionality with powerful simulation capabilities.
 {% endhint %}
 
+{% hint style="info" %}
+**✨ New: Sub-Second Latency with Partial Blocks**
+
+Tycho now provides early support for partial blocks on Base, enabling sub-second latency by streaming pre-confirmation updates. Enable via `--partial-blocks` (CLI), `partial_blocks=True` (Python), or `.enable_partial_blocks()` (Rust). See [Streaming Options](./#streaming-options) below for details.
+{% endhint %}
+
 ### Key Features
 
 * **Real-Time Streaming**: Get low-latency updates to stay in sync with the latest protocol changes. Discover new pools as they’re created.
@@ -72,6 +78,50 @@ tycho-client --remove-tvl-threshold 95 --add-tvl-threshold 100 --exchange uniswa
 
 This will stream state updates for all components whose TVL exceeds the `add-tvl-threshold`. It will continue to track already added components if they drop below the `add-tvl-threshold`, only emitting a message to remove them if they drop below `remove-tvl-threshold`.
 
+#### Streaming Options
+
+Tycho Client supports several options to customize the data stream. These are available as CLI flags, Rust builder methods, and Python parameters. Refer to each client's documentation for usage details.
+
+| Option                  | Description                                        | When to use                     |
+| ----------------------- | -------------------------------------------------- | ------------------------------- |
+| **Partial blocks**      | Stream pre-confirmation blocks                     | For lower stream latency        |
+| **No state**            | Stream only component metadata and tokens          | For lower stream latency        |
+| **Include TVL**         | Attach approximate TVL estimates to each component | For getting components TVL      |
+| **No TLS**              | Use unencrypted transports (http/ws)               | For local self-hosted indexers  |
+| **Disable compression** | Turn off stream message compression                | _Debugging Only_                |
+
+{% hint style="info" %}
+**Note:** `disable compression` and `no tls` are not intended to be used with the [hosted](../../hosted-endpoints.md) Tycho Indexer endpoints.
+{% endhint %}
+
+<details>
+
+<summary><strong>Details:</strong> <strong>Partial Blocks</strong></summary>
+
+Some chains, such as [Base](https://docs.base.org/building-with-base/differences/flashblocks), support _flash blocks_ - pre-confirmation updates that contain parts of a future block before its construction is finished. When `partial blocks` is enabled, Tycho streams these incremental updates as they arrive, giving you sub-block latency. On chains without flash block support, enabling this flag is unsupported.
+
+{% hint style="warning" %}
+Block hashes in partial block messages are **unstable** — they change between partial updates and will differ from the final block hash. Do not use them as persistent identifiers or cache keys. See the [Substreams documentation](https://docs.substreams.dev/reference-material/chain-support/flashblocks#developing-for-partial-blocks) for details.
+{% endhint %}
+
+</details>
+
+<details>
+
+<summary><strong>Details: No State</strong></summary>
+
+By default, the first sync message includes full component snapshots, and every subsequent block includes state deltas (reserves, balances, contract storage). If you only need to discover which components exist and which tokens they involve, such as to build a pool registry or monitor new deployments, you can disable state monitoring with `no state`. This significantly reduces message sizes, startup time, and processing overhead, as both snapshots and per-block state updates are omitted entirely.
+
+</details>
+
+<details>
+
+<summary><strong>Details: Include TVL</strong></summary>
+
+When enabled, each message includes an approximate TVL estimate for every tracked component. This is useful for building dashboards or for ranking pools by liquidity. Note that enabling this option increases startup latency: for each snapshot request, the client makes additional RPC calls to fetch token prices and compute TVL for all tracked components. This overhead scales with the number of components you're tracking.
+
+</details>
+
 ***
 
 ## Understanding Tycho Client Messages

docs/for-solvers/simulation.md

@@ -10,6 +10,12 @@ Tycho Simulation is a Rust crate that provides powerful tools for **interacting
 
 The repository is available [here](https://github.com/propeller-heads/tycho-simulation).
 
+{% hint style="info" %}
+**✨ New: Sub-Second Latency with Partial Blocks**
+
+Tycho now provides early support for partial blocks on Base, enabling sub-second latency by streaming pre-confirmation updates. Enable this feature with `.enable_partial_blocks()` on your `ProtocolStreamBuilder`.
+{% endhint %}
+
 ## Installation
 
 The `tycho-simulation` package is available on [Github](https://github.com/propeller-heads/tycho-simulation).
@@ -210,6 +216,10 @@ The first message received will contain states for all protocol components regis
 
 For a full list of supported protocols and the simulation state implementations they use, see [Supported Protocols](supported-protocols.md).
 
+{% hint style="info" %}
+`ProtocolStreamBuilder` supports the same [streaming options](https://docs.propellerheads.xyz/tycho/for-solvers/indexer/tycho-client#streaming-options) as the Tycho Client, with one difference: TVL estimates are **always included** in the simulation package and cannot be disabled.
+{% endhint %}
+
 <details>
 
 <summary>Example: Consuming the Stream and Simulating</summary>