> For the complete documentation index, see [llms.txt](https://docs.sigrex.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.sigrex.io/more/changelog.md). # Changelog ## `0.4.1` Added [**Outcome.xyz**](https://outcome.xyz/) as a supported prediction market for Prediction Signal Bots, expanding prediction market trading integrations. Code Strategies can now be created **without Signal Generation settings**, making it possible to use them as lightweight runtimes without unnecessary configuration fields. Added a default `Content-Type: application/json` header to the `$.Http.post()` and `$.Http.put()` methods in Code Strategies. The header no longer needs to be specified manually when making JSON requests. Added the **`get_candlestick_data`** tool to LLM Strategies, allowing them to retrieve historical candlestick data directly for analysis. Introduced a new **LLM service call timeout policy** based on the LLM Session cron interval. The timeout is automatically adjusted according to the session frequency: | LLM Session Interval | Timeout | | -------------------- | ----------- | | 1 minute | 30 seconds | | 3 minutes | 1.5 minutes | | 5 minutes | 2.5 minutes | | 15 minutes | 7.5 minutes | | 30 minutes | 7.5 minutes | | 1 hour | 7.5 minutes | | 2 hours | 7.5 minutes | | 4 hours | 7.5 minutes | | 8 hours | 7.5 minutes | | 12 hours | 7.5 minutes | | 1 day | 7.5 minutes | Increased the **LLM Reaction service call timeout** to **5 minutes**, allowing longer-running reactions to complete successfully. Also included several **UI improvements and refinements** across the platform. ## `0.4.0` Added [**SiliconFlow** ](https://cloud.siliconflow.com/)and [**SambaNova** ](https://sambanova.ai/)as new LLM providers for LLM Strategies, giving strategies access to even more AI models and inference options. LLM Strategies can now **start and stop themselves or other strategies**, retrieve their own or other strategies' configurations, and modify strategy logic directly. LLMs can update the **`code`** field of Code Strategies and the **`prompt`** field of LLM Strategies, enabling new possibilities for self-improving, self-reviewing, and strategy-management workflows. Added `$.Strategy.id` and `$.Strategy.type` to the Code Strategy runtime. These provide the strategy's unique identifier and type (`code-strategy` or `code-reaction`). Added `$.getExchangeRate()` to Code Strategies, making it easier to retrieve the current price of a symbol from any exchange supported by Sigrex's live exchange-rate system. > **Note:** The injected `$.Price` object is deprecated and will be removed in a future release. Use `$.getExchangeRate()` instead. Added `$.Strategy.stop()` to Code Strategies. This sets the strategy status to `INACTIVE` and prevents future executions. It does **not** immediately terminate the currently running process. To stop the current execution as well, use `return` after calling `$.Strategy.stop()`. Prediction Signal Bots now support **Wildcard / Slug from Signal** mode. Instead of being restricted to a single market slug, bots can dynamically process market slugs provided in the incoming signal payload. Because the market slug must be verified before each order, this mode is slightly slower than using a fixed slug. For bots that only trade a specific market, using a manually configured slug is recommended. Introduced an [**AI Agent Skill Library**](https://github.com/sigrexio/skills), providing reusable skills and capabilities for building more powerful AI agents and automation workflows. ## `0.3.4` LLM Strategies now support [**OpenRouter**](https://openrouter.ai/), offering access to a wide range of AI models through a single provider, along with [Groq](https://groq.com/) and [Cerebras](https://www.cerebras.ai/) for high-performance inference. Added native [`web_fetch`](https://docs.sigrex.io/startegies/llm-session/signal-generation#web_fetch) and [`execute_javascript`](https://docs.sigrex.io/startegies/llm-session/signal-generation#execute_javascript) tools to LLM Strategies. `web_fetch` allows AI strategies to retrieve external data on demand without requiring every URL to be predefined in the prompt. `execute_javascript` runs JavaScript inside the same secure sandbox used by [Code Strategies](/startegies/code.md), enabling AI strategies to perform calculations, process data, and interact with external APIs. > **Note:** The `{{get:}}` template variable remains the most cost-efficient and configurable option for predefined HTTP requests. The `web_fetch` tool is intended for dynamic requests determined by the AI during execution. Added the [`send_email`](https://docs.sigrex.io/startegies/llm-session/signal-generation#send_email) tool for LLM Strategies. AI strategies can now send emails directly when prompted, making it possible to deliver custom alerts, reports, notifications, or other important messages whenever a specific condition or workflow requires user attention. Added [`get_storage`](https://docs.sigrex.io/startegies/llm-session/signal-generation#get_storage) and [`append_storage`](https://docs.sigrex.io/startegies/llm-session/signal-generation#append_storage) tools for LLM Strategies. AI strategies can now retrieve their persistent storage or append new data without replacing the existing contents. The maximum storage size has also been increased to **25 MB**. Improved the [`get_symbol_price`](https://docs.sigrex.io/startegies/llm-session/signal-generation#get_symbol_price) tool for LLM Strategies. AI strategies can now retrieve multiple exchange rates in a single request, with support for up to **20 currency pairs per call**, reducing unnecessary API requests and improving execution efficiency. Added [`$.Http.createApiClient()`](https://docs.sigrex.io/startegies/code#usd.http) to Code Strategies and Code Reactions. This helper creates a fully configured [Sigrex API client](https://github.com/sigrexio/sigrex-client-ts), making it easier to interact with the Sigrex API directly from strategy code. ## `0.3.3` A major frontend overhaul with numerous UI and UX improvements, delivering a faster, cleaner, and more intuitive experience throughout the platform. Added support for [**Nous Research**](https://nousresearch.com/) and [**NVIDIA**](https://build.nvidia.com/), giving you access to even more state-of-the-art language models for your AI strategies. Reactions now support **Frequency** settings, allowing you to control how often a reaction can be triggered and helping prevent unwanted repeated executions. ## `0.3.2` All bots and strategies can now be organized into folders, making it easier to manage and structure your workspace as it grows. [Code Strategies](/startegies/code.md) and [Reactions](/reaction/code-reaction.md) now support folder-level environment variables. Variables defined on a folder automatically apply to all strategies within that folder and will override globally defined environment variables when the same key is used. Restricted platform access as part of our ongoing security and compliance measures. Services are currently unavailable connecting from: * Afghanistan * Belarus * Bosnia and Herzegovina * Guinea * Guinea-Bissau * Iran * Iraq * Russia ## `0.3.1` Added **ML-DSA44 post-quantum signing** as a new API authentication option alongside HMAC SHA256. You can now secure requests using either: * HMAC * ML-DSA44 ([post-quantum safe](https://openquantumsafe.org/liboqs/algorithms/sig/ml-dsa.html)) This brings NIST-standard quantum-resistant signatures to the developer API. ## `0.3.0` LLM Sessions no longer rely on parsing plain-text responses such as `LONG`, `SHORT`, or `EXIT`. Signal generation has been upgraded to native tool calling, allowing AI strategies to open and close positions directly through built-in trading tools. Introduced the new `set_storage` [tool](/startegies/llm-session/signal-generation.md#llm-tooling-reference). LLM strategies can now maintain their own persistent JSON storage, which is automatically provided on every execution. This enables AI strategies to track state, remember previous decisions, manage workflows, and build more advanced long-term reasoning processes. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://docs.sigrex.io/more/changelog.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.