llm.rb 11.3.1 → 12.0.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: e3aad41c57654e1abef4c184da8904ae5840708e8f5ce937e1030279a5e5a240
-  data.tar.gz: 9b03ac6a4f04190a57f2e4698eab923087febbab6a1ff8ad67f391fce4bd5c1e
+  metadata.gz: e042f679a5f682feed6726430dd1b28fd6728c9223cbd0e0499c92e8952a06e6
+  data.tar.gz: 0cb3f55e70a321bea3f1903261142bf3aa842fface494331a548f684aac520d8
 SHA512:
-  metadata.gz: f429a75d5b612eb27968b6e6451bd231f2191b5b955262503c68baa7ca27f289c01ad564558e82d7fcc732393f635c32caf883a36815b41d86f6c1289cfffb61
-  data.tar.gz: 90924b997b5d47d74c3df6e303c890fb5d1a771bae10e63c349436bcbd061241604184f62df87cb2dee7d85405fbd53f08d279ef6b4d3a85a12b0d59b7377229
+  metadata.gz: ba64ba9b87d1343f9d94a06b97cb6cc0ec0ca5d7a8c4b235e52c9ed489c5670848b44e0549e838b01b12be89d19005ba571a6dd9557a8f2f5f4e62a3ec018ec9
+  data.tar.gz: 81a489afa510edf608b74a3a83abc16b6d81c6f59f614df6460f638dbcf1e0a539d8d2b6b2e5a98a3772d5c9d7c7233d7b231bbf447dda7cfcf8df1861819c8d

data/CHANGELOG.md

@@ -1,7 +1,248 @@
-# Changelog
+<p align="center">
+  <a href="https://r.uby.dev">
+    <img
+      src="https://github.com/r-uby-dev/llm.rb/raw/main/rubydev.svg"
+      width="400"
+      height="200"
+      border="0"
+      alt="a r.uby.dev project"
+     >
+  </a>
+</p>
+
+> Changelog <br>
+> a [r.uby.dev](https://r.uby.dev) project
 
 ## Unreleased
 
+## v12.0.0
+
+Changes since `v11.3.1`.
+
+This release relicenses the project under the Business Source License,
+defaults OpenAI to the Responses API and gpt-image models, adds the
+DeepInfra provider with audio and image support, introduces
+DeepSeek vector-graphics generation and schema support, extends xAI
+image editing, adds `LLM::Schema.defaults` and schema string rendering,
+and makes ActiveRecord and Sequel agent wrappers yield `LLM::Agent`
+instead of polluting the model namespace.
+
+### Breaking
+
+* **License change** <br>
+  The llm.rb runtime has been developed primarily by one
+  person for 3 years. That was done on my own time, and
+  I haven't made a dime from that work.
+
+  So when I saw a multi-million dollar company benefit from
+  the work and for it to become the backbone of their AI
+  infrastructure and then see them not contribute back or
+  offer any kind of support, I decided this is not sustainable,
+  or fair.
+
+  I assumed good faith and for people to act in the spirit of
+  open source but sadly, that's just not the case. I
+  have to choose a license that respects my time and effort.
+
+  For those reasons, llm.rb is being relicensed under the
+  [Business Source license](https://mariadb.com/bsl11/).
+  So what does that mean?
+
+  In a nutshell:
+
+  * Free for personal use.
+  * Free for education.
+  * Free for evaluation, development, and testing.
+  * Commercial production use requires a commercial license.
+  * Exemptions on a case-by-case basis
+
+  After 4 years, the license expires and it will become
+  available under the 0BSDL as it was before v12.0.0.
+  These 4 years apply to a specific version, and not the
+  project overall.
+
+  Going forward, v12.0.0 will be relicensed to respect
+  my time, energy, and effort. llm.rb took an incredible
+  amount of time and effort, and continues to do so, so
+  I want to protect myself from companies who benefit
+  from my work but don't respect the time or effort that
+  was put into it.
+
+* **OpenAI: default to the Responses API** <br>
+  The responses API has both models and features that are unavailable
+  on the chat completions API, and the responses API appears to be
+  the API of the future for OpenAI.
+
+  Worth noting: the llm.rb implementation does **not** store state
+  server-side by default. This can be changed with the `store: true`
+  option. The legacy chat completions API can be accessed with the
+  `mode: :completions` option.
+
+  llm.rb has had support for the responses API for quite
+  a while but it was not the default, and a number of bugs
+  were found and fixed during the process of making it the
+  default.
+
+* **OpenAI: use gpt-image for image generation** <br>
+  The `dalle` models are in the process of being deprecated, and support
+  has been dropped from llm.rb. The `gpt-image` models are the next-generation
+  image-generation models from OpenAI.
+
+* **xAI: provide images as base64-encoded data** <br>
+  Both xAI, and OpenAI had the option to generate images via a URL
+  you can fetch, or as a base64-encoded string embedded directly
+  in the response.
+
+  OpenAI is moving away from the URL transport since deprecating dalle,
+  and with that in mind, llm.rb has dropped support for the URL transport
+  across all providers that supported it.
+
+  Google, xAI, and OpenAI now consistently provide generated and modified
+  images as a base64-encoded string.
+
+* **ActiveRecord: yield `LLM::Agent` to `acts_as_agent`** <br>
+  With this change we yield an instance of `LLM::Agent` to the `acts_as_agent`
+  method, and drop the methods (such as `model`, `instructions`, etc) that
+  were previously defined directly on the model. This keeps the number of
+  methods that llm.rb adds to an ActiveRecord model at a minimum and retains
+  the same capabilities as before.
+
+* **Sequel: yield `LLM::Agent` to `plugin(:agent)`** <br>
+  Ditto as above but for Sequel.
+
+* **Remove the langsmith tracer** <br>
+  This code was contributed by a third party but contains
+  many anti-patterns that are against llm.rb conventions
+  and best practices. It was merged without oversight or
+  review, and basically against the ethos of open source.
+
+  I also don't have a langsmith account to maintain the
+  code. The alternative is the `LLM::Tracer::Telemetry` class
+  that was originally written by me, and serves as a
+  general-purpose OTP tracer.
+
+### Add
+
+* **Add a new provider: LLM::DeepInfra** <br>
+  [DeepInfra](https://deepinfra.com) provide OpenAI-compatible
+  endpoints for a large catalog of hosted open-source and
+  open-weight models. <br> Capabilities like tool calling, structured outputs, and
+  reasoning can depend on the model.
+
+* **Add new image provider: LLM::DeepInfra::Images** <br>
+  [DeepInfra](https://deepinfra.com) provide access to
+  diverse set of text-to-image models. <br> Learn more about the
+  available models on their [text-to-image models](https://deepinfra.com/models/text-to-image)
+  page.
+
+* **DeepSeek: add `LLM::DeepSeek::Images#create` and `#edit`** <br>
+  This new API can generate and edit vector graphics (SVGs). <br>
+  It is an experimental approach and API.
+
+  DeepSeek does not provide an image generation model however
+  its text-to-text models can generate SVG documents, and
+  that's the approach this feature takes. It is limited
+  to vector graphics rather than raster images.
+
+* **DeepSeek: attach `LLM::Response#agent` to image responses** <br>
+  The DeepSeek image API is built on top of
+  [`LLM::Agent`](https://r.uby.dev/api-docs/llm.rb/LLM/Agent.html).
+  Image responses now expose that agent via `res.agent`, which makes
+  it possible to carry the same session across multiple generations
+  or edits.
+
+* **xAI: add `LLM::XAI::Images#edit`** <br>
+  With this change it is possible to both generate images
+  from a prompt, and edit an existing image with a prompt.
+  xAI now has the same edit and create capabilities that
+  OpenAI has.
+
+* **Add `LLM::Schema.defaults`** <br>
+  This method lets you map multiple property names to
+  different default values. It is similar to `LLM::Schema.required`
+  in the sense that it is called after the properties of
+  a schema have been defined.
+
+* **Add `LLM::Schema#to_s` and `LLM::Schema.to_s`** <br>
+  Schemas can now be rendered as a prompt-friendly string.
+  This is useful when the shape of a schema needs to be
+  described in natural-language instructions rather than
+  passed through a native structured output interface.
+
+* **DeepSeek: add `LLM::Schema` support** <br>
+  DeepSeek can now use `schema:` for structured output.
+  llm.rb handles this by setting `response_format: {type: "json_object"}`
+  and describing the schema in a system message.
+
+* **OpenAI: add local file support to the Responses API** <br>
+  Our responses API implementation lacked local file support. <br>
+  This change fixes that by supporting both image, document,
+  and other media types that OpenAI may support.
+
+* **Add `LLM::Response#id` across all providers** <br>
+  This method was previously implemented via `method_missing`,
+  and the field name could change depending on the provider.
+  The new method is a catch-all that provides a single method
+  that works across all providers.
+
+* **Add `LLM::DeepInfra::Audio`** <br>
+  DeepInfra implements most of the llm.rb audio interface
+  with both the `create_speech` and `create_transcription`
+  methods. The `create_translation` method is not implemented,
+  and the available text-to-speech and speech-to-text models
+  are more varied than other providers.
+
+* **OpenAI: normalize text-to-speech responses** <br>
+  The `res.audio` method now returns an
+  [`LLM::URIData`](https://r.uby.dev/api-docs/llm.rb/LLM/URIData.html)
+  object for OpenAI text-to-speech responses. The object provides
+  `encoded`, `decoded`, `content_type`, and `encoding_type`.
+
+* **DeepInfra: normalize text-to-speech responses** <br>
+  The `res.audio` method now returns an
+  [`LLM::URIData`](https://r.uby.dev/api-docs/llm.rb/LLM/URIData.html)
+  object for DeepInfra text-to-speech responses. The object provides
+  `encoded`, `decoded`, `content_type`, and `encoding_type`.
+
+### Fix
+
+* **Fix Google `temperature` parameter fall-through** <br>
+  Ensure provider-level `temperature` and other `generationConfig`
+  parameters are forwarded to the API correctly instead of being
+  silently dropped.
+
+* **Fix Google `generationConfig` collisions** <br>
+  Prevent duplicate or conflicting `generationConfig` keys in the
+  Google request adapter.
+
+### Change
+
+* **Change OpenAI defaults** <br>
+  The default chat model is now `gpt-5.4-mini`. <br>
+  The default image model is now `gpt-image`.
+
+* **Change google defaults** <br>
+  The default chat model is now `gemini-3.1-flash-lite` <br>
+  The default embeddings model is now `gemini-embedding-2`
+
+* **Change xAI defaults** <br>
+  The default chat model is now `grok-4.3`. <br>
+  The default image model is now `grok-imagine-image-quality`.
+
+* **Return an `LLM::Object` from `LLM::Response#content!`** <br>
+  The Hash-like, indifferent access data structure known as
+  `LLM::Object` provides a convenient interface around a Hash
+  object. It allows method access via `obj.key`, and decays
+  into a Hash in many cases.
+
+  The `LLM::Response#content!` method now wraps its content
+  in an `LLM::Object` but only after it has parsed its
+  content (a JSON string) into a Ruby data structure.
+
+* **Refresh model metadata** <br>
+  Update `data/*.json` files with current provider model listings,
+  pricing, and capabilities.
+
 ## v11.3.1
 
 Changes since `v11.3.0`.

data/LICENSE

@@ -1,17 +1,92 @@
-Copyright (C) 2026
-Antar Azri <azantar@proton.me>
-0x1eef <0x1eef@hardenedbsd.org>
-
-Permission to use, copy, modify, and/or distribute this
-software for any purpose with or without fee is hereby
-granted.
-
-THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS
-ALL WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL
-IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO
-EVENT SHALL THE AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT,
-INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER
-RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
-ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION,
-ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE
-OF THIS SOFTWARE.
+Business Source License 1.1
+
+License text copyright (c) 2024 MariaDB plc, All Rights Reserved.
+"Business Source License" is a trademark of MariaDB plc.
+
+Parameters
+
+Licensor: Robert Gleeson
+
+Licensed Work: llm.rb
+
+Additional Use Grant:
+Free for personal use.
+Free for education.
+Free for evaluation, development, and testing.
+Commercial production use requires a commercial license.
+
+Change Date:
+Four years after the first public release of each specific version of the
+Licensed Work.
+
+Change License:
+BSD Zero Clause License (0BSD)
+
+Terms
+
+The Licensor hereby grants you the right to copy, modify, create derivative
+works, redistribute, and make non-production use of the Licensed Work. The
+Licensor may make an Additional Use Grant, above, permitting limited
+production use.
+
+Effective on the Change Date, or the fourth anniversary of the first publicly
+available distribution of a specific version of the Licensed Work under this
+License, whichever comes first, the Licensor hereby grants you rights under
+the terms of the Change License, and the rights granted in the paragraph
+above terminate.
+
+If your use of the Licensed Work does not comply with the requirements
+currently in effect as described in this License, you must purchase a
+commercial license from the Licensor, its affiliated entities, or authorized
+resellers, or you must refrain from using the Licensed Work.
+
+All copies of the original and modified Licensed Work, and derivative works
+of the Licensed Work, are subject to this License. This License applies
+separately for each version of the Licensed Work and the Change Date may vary
+for each version of the Licensed Work released by Licensor.
+
+You must conspicuously display this License on each original or modified copy
+of the Licensed Work. If you receive the Licensed Work in original or
+modified form from a third party, the terms and conditions set forth in this
+License apply to your use of that work.
+
+Any use of the Licensed Work in violation of this License will automatically
+terminate your rights under this License for the current and all other
+versions of the Licensed Work.
+
+This License does not grant you any right in any trademark or logo of
+Licensor or its affiliates (provided that you may use a trademark or logo of
+Licensor as expressly required by this License).
+
+TO THE EXTENT PERMITTED BY APPLICABLE LAW, THE LICENSED WORK IS PROVIDED ON
+AN "AS IS" BASIS. LICENSOR HEREBY DISCLAIMS ALL WARRANTIES AND CONDITIONS,
+EXPRESS OR IMPLIED, INCLUDING (WITHOUT LIMITATION) WARRANTIES OF
+MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, NON-INFRINGEMENT, AND
+TITLE.
+
+MariaDB hereby grants you permission to use this License's text to license
+your works, and to refer to it using the trademark "Business Source
+License", as long as you comply with the Covenants of Licensor below.
+
+Covenants of Licensor
+
+In consideration of the right to use this License's text and the "Business
+Source License" name and trademark, Licensor covenants to MariaDB, and to all
+other recipients of the licensed work to be provided by Licensor:
+
+To specify as the Change License the GPL Version 2.0 or any later version, or
+a license that is compatible with GPL Version 2.0 or a later version, where
+"compatible" means that software provided under the Change License can be
+included in a program with software provided under GPL Version 2.0 or a later
+version. Licensor may specify additional Change Licenses without limitation.
+
+To either: (a) specify an additional grant of rights to use that does not
+impose any additional restriction on the right granted in this License, as
+the Additional Use Grant; or (b) insert the text "None" to specify a Change
+Date. Not to modify this License in any other way.
+
+Notice
+
+The Business Source License (this document, or the "License") is not an Open
+Source license. However, the Licensed Work will eventually be made available
+under an Open Source License, as stated in this License.