ollama_dart

Adds renderer and parser to CreateRequest, matching the upstream OpenAPI spec's new tool-calling prompt renderer (e.g. qwen3.5) and output parser (e.g. harmony) fields. Also adds client-ahead fields for full api/types.go parity that the spec still omits — files/adapters (filename → SHA256 digest maps, per the Create a Model docs) for creating models from GGUF/Safetensors and LoRA adapters, plus draftQuantize/draftFiles, remoteHost, requires, and info. Also fixes CreateRequest's ==/hashCode/toString contract, which previously only compared model, to cover all 18 fields with content-based map/list equality.

• FEAT: CreateRequest renderer/parser and files/adapters (#285). (c8a61de9)

Adds the new max thinking level (ThinkLevel.max), matching Ollama's spec update that added "max" to the think enum on ChatRequest and GenerateRequest — use ChatRequest(..., think: ThinkWithLevel(ThinkLevel.max)) to request the highest thinking level on supported reasoning models. Also fixes buildUrl producing a double slash (//api/...) when OllamaConfig.baseUrl (or OLLAMA_HOST) has a trailing slash, and preserves base-URL query params (including repeated keys) that were previously dropped — proxy setups with tokens in the base URL now work correctly.

• FEAT: Add "max" thinking level (#267). (8061f857)
• FIX: Avoid double slash and preserve base URL query params in buildUrl (#273). (7686e397)

Adds experimental image generation to the /api/generate endpoint: GenerateRequest gains width/height/steps, and GenerateResponse and the streaming GenerateStreamEvent gain image (base64), completed, and total for diffusion progress. These fields are implemented client-side ahead of the upstream spec, which keeps them out of openapi.yaml while they remain experimental. Also adds the draftNumPredict model option to ModelOptions and completes the previously-partial ==/hashCode/toString contracts on GenerateStreamEvent and ChatStreamEvent, which had compared only a subset of fields.

• FEAT: Add experimental image generation and draft_num_predict option (#254). (04b901a3)

Fixes browser usage (Flutter Web / dart2wasm) by no longer sending the non-CORS-safelisted X-Request-ID header by default — Ollama's Access-Control-Allow-Headers list excludes it, so the preflight previously failed and blocked all browser requests against a real Ollama server. A request ID is still generated internally for logging, error correlation, and retry/abort tracing; the new OllamaConfig(sendRequestIdHeader: true) opt-in restores the wire header for callers behind a proxy configured to accept it, and an X-Request-ID set explicitly via defaultHeaders is always sent.

• FEAT: Make X-Request-ID header opt-in to fix browser CORS (#228). (f497a722)

Annotates llms.txt with per-link token counts and per-package totals so coding agents can budget context before fetching documentation, examples, or changelogs — inspired by Addy Osmani's Agentic Engine Optimization article.

• FEAT: Annotate llms.txt with token counts and tighten agent-facing docs (#181). (a1e82aca)

Adds a strict semver bullet to the package README.

• DOCS: Overhaul root README and add semver bullet to all packages (#151). (e6af33dd)

[!CAUTION] This release has breaking changes. See the Migration Guide for upgrade instructions.

Replaces untyped Object/Object? fields with sealed union types (KeepAlive, EmbedInput, StopSequence) for improved type safety across request models. Also adds a name field to RunningModel from the upstream Ollama API update and overhauls README documentation with llms.txt ecosystem files.

• BREAKING REFACTOR: Replace Object fields with sealed union types (#120). (5d7a4f24)
• FEAT: Add name field to RunningModel and update spec (#142). (4fc2e1bb)
• DOCS: Overhaul READMEs and add llms.txt ecosystem (#149). (98f11483)

Fixed verification warnings in generated model classes.

• FIX: Resolve verification warnings (#98). (da231740)

This release adds inline streaming error detection for improved reliability when handling streamed responses.

• FEAT: Detect inline streaming errors (#91). (9f0eaf37)
• DOCS: Improve READMEs with badges, sponsor section, and vertex_ai deprecation (#90). (5741f2f3)

Added missing fields to chat, completion, and model response models.

• FEAT: Add missing fields to response models (#80). (73e44241)

Internal improvements to build tooling and package publishing configuration.

• REFACTOR: Migrate API skills to the shared api-toolkit CLI (#74). (923cc83e)
• CHORE: Add .pubignore to exclude .agents/ and specs/ from publishing (#78). (0ff199bf)

Added baseUrl and defaultHeaders parameters to withApiKey constructors, fixed hashCode for list fields, and unified equality helpers.

• FEAT: Add baseUrl and defaultHeaders to withApiKey constructors (#57). (f0dd0caa)
• FIX: Use Object.hashAll() for list fields in hashCode (#65). (4b19abd9)
• REFACTOR: Unify equality_helpers.dart across packages (#67). (ec2897f8)

Added withApiKey convenience constructor for simplified client initialization.

• FEAT: Add withApiKey convenience constructors (#56). (b06e3df3)
• CHORE: Bump googleapis from 15.0.0 to 16.0.0 and Dart SDK to 3.9.0 (#52). (eae130b7)
• CI: Add GitHub Actions test workflow (#50). (6c5f079a)

[!CAUTION] This release has breaking changes. See the Migration Guide for upgrade instructions.

TL;DR: Complete reimplementation with a new architecture, minimal dependencies, resource-based API, and improved developer experience. Hand-crafted models (no code generation), interceptor-driven architecture, comprehensive error handling, and full Ollama API coverage.

What's new

• Resource-based API organization:
  • client.chat — Chat completions (multi-turn conversations)
  • client.completions — Text generation (single-turn)
  • client.embeddings — Generate text embeddings
  • client.models — Model management (list, show, pull, push, copy, delete, create, ps)
  • client.version — Server version info
• Architecture:
  • Interceptor chain (Auth → Logging → Error → Transport with Retry wrapper).
  • Authentication: Bearer token or custom via AuthProvider interface.
  • Retry with exponential backoff + jitter (only for idempotent methods on 429, 5xx, timeouts).
  • Abortable requests via abortTrigger parameter.
  • NDJSON streaming parser for real-time responses.
  • Central OllamaConfig (timeouts, retry policy, log level, baseUrl, auth).
• Hand-crafted models:
  • No code generation dependencies (no freezed, json_serializable).
  • Minimal runtime dependencies (http, logging only).
  • Immutable models with copyWith using sentinel pattern.
  • Full type safety with sealed exception hierarchy.
• Type-safe enums and sealed classes:
  • DoneReason enum for completion stop reasons (stop, length, load, unload).
  • ThinkValue sealed class for thinking mode (ThinkEnabled, ThinkWithLevel).
  • ResponseFormat sealed class for format options (JsonFormat, SchemaFormat).
  • MessageRole enum for message roles (system, user, assistant, tool).
• Improved DX:
  • Simplified model names (e.g., ChatRequest instead of GenerateChatCompletionRequest).
  • Named constructors for common patterns (e.g., ChatMessage.user(), ChatMessage.system()).
  • Explicit streaming methods (createStream() vs create()).
  • Rich logging with field redaction for sensitive data.
• Full API coverage:
  • Chat completions with tool calling support.
  • Text completions with thinking mode.
  • Embeddings generation.
  • Model management (list, show, create, copy, delete, pull, push, ps).
  • Server version info.

Breaking Changes

• Resource-based API: Methods reorganized under strongly-typed resources:
  • client.generateChatCompletion() → client.chat.create()
  • client.generateChatCompletionStream() → client.chat.createStream()
  • client.generateCompletion() → client.completions.generate()
  • client.generateCompletionStream() → client.completions.generateStream()
  • client.generateEmbedding() → client.embeddings.create()
  • client.listModels() → client.models.list()
  • client.showModelInfo() → client.models.show()
  • client.listRunningModels() → client.models.ps()
  • client.getVersion() → client.version.get()
• Model class renames:
  • GenerateChatCompletionRequest → ChatRequest
  • GenerateChatCompletionResponse → ChatResponse
  • GenerateCompletionRequest → GenerateRequest
  • GenerateCompletionResponse → GenerateResponse
  • GenerateEmbeddingRequest → EmbedRequest
  • GenerateEmbeddingResponse → EmbedResponse
  • Message → ChatMessage
  • Tool → ToolDefinition
  • RequestOptions → ModelOptions
• Configuration: New OllamaConfig with AuthProvider pattern.
• Exceptions: Replaced OllamaClientException with typed hierarchy:
  • ApiException, ValidationException, RateLimitException, TimeoutException, AbortedException.
• Dependencies: Removed freezed, json_serializable; now minimal (http, logging).
• Base URL: No longer needs /api suffix — just use http://localhost:11434.

See MIGRATION.md for step-by-step examples and mapping tables.

Commits

• BREAKING FEAT: Complete v1.0.0 reimplementation (#1). (0ff41032)
• FEAT: Add type-safe enums and sealed classes (#22). (73372581)
• FIX: Pre-release documentation and code fixes (#45). (b33ae6d5)
• REFACTOR: Align client package architecture across SDK packages (#37). (cf741ee1)
• REFACTOR: Align API surface across all SDK packages (#36). (ed969cc7)
• REFACTOR: Extract streaming helpers to StreamingResource mixin (#2). (0b6f0ed9)
• DOCS: Refactors repository URLs to new location. (76835268)

[!CAUTION] This release has breaking changes. See the Migration Guide for upgrade instructions.

• FEAT: Enhance CreateModelRequest with new fields (#802). (c5c73549)
• FEAT: Add tool_name and index support (#800). (f0f77286)
• FEAT: Add remote_model and remote_host support (#799). (36b9d5f2)
• FEAT: Add truncate and shift support (#798). (098a0815)
• FEAT: Support high, medium, low for think (#797). (1cbe3fcf)
• FEAT: Support JSON schema in ResponseFormat (#796). (2f399465)
• FEAT: Upgrade to http v1.5.0 (#785). (f7c87790)
• BREAKING REFACTOR: Improve factory names (#806). (fbfa7acb)
• BREAKING BUILD: Require Dart >=3.8.0 (#792). (b887f5c6)

• REFACTOR: Fix pub format warnings (#809). (640cdefb)

• FEAT: Migrate to Freezed v3 (#773). (f87c8c03)

• BUILD: Update dependencies (#751). (250a3c6)

• FEAT: Add think/thinking params to ollama_dart (#721). (701d7968)
• FEAT: Add capabilities, projector_info, tensors and modified_at to Ollama's ModelInfo (#690). (c5e247db)
• FEAT: Update dependencies (requires Dart 3.6.0) (#709). (9e3467f7)
• REFACTOR: Remove fetch_client dependency in favor of http v1.3.0 (#659). (0e0a685c)
• FIX: Fix linter issues (#656). (88a79c65)

• FEAT: Update Ollama default model to llama-3.2 (#554). (f42ed0f0)

• REFACTOR: Add new lint rules and fix issues (#621). (60b10e00)
• REFACTOR: Upgrade api clients generator version (#610). (0c8750e8)

• FEAT: Add support for min_p in Ollama (#512). (e40d54b2)

[!CAUTION] This release has breaking changes. See the Migration Guide for upgrade instructions.

• FEAT: Add tool calling support (#504). (1ffdb41b)
• BREAKING FEAT: Update Ollama default model to llama-3.1 (#506). (b1134bf1)
• FEAT: Add support for Ollama version and model info (#488). (a110ecb7)
• FEAT: Add suffix support in Ollama completions API (#503). (30d05a69)
• BREAKING REFACTOR: Change Ollama push model status type from enum to String (#489). (90c9ccd9)
• DOCS: Update Ollama request options default values in API docs (#479). (e1f93366)

• FEAT: Add support for listing running Ollama models (#451). (cfaa31fb)
• REFACTOR: Migrate conditional imports to js_interop (#453). (a6a78cfe)

• FEAT: Support buffered stream responses (#445). (ce2ef30c)
• FIX: Fix deserialization of sealed classes (#435). (7b9cf223)

[!CAUTION] This release has breaking changes. See the Migration Guide for upgrade instructions.

• BREAKING FEAT: Align Ollama client to the Ollama v0.1.36 API (#411). (326212ce)
• FEAT: Update Ollama default model from llama2 to llama3 (#417). (9d30b1a1)
• FEAT: Add support for done reason (#413). (cc5b1b02)

• FIX: digest path param in Ollama blob endpoints (#430). (2e9e935a)

• FEAT: Add Ollama keep_alive param to control how long models stay loaded (#319). (3b86e227)
• FEAT: Update meta and test dependencies (#331). (912370ee)
• DOCS: Update pubspecs. (d23ed89a)

• FIX: Have the == implementation use Object instead of dynamic (#334). (89f7b0b9)

• FEAT: Add support for chat API and multi-modal LLMs (#274). (76e1a294)

• DOCS: Update CHANGELOG.md. (d0d46534)

• FEAT: Implement ollama_dart, a Dart client for Ollama API (#238). (d213aa9c)

• DOCS: Update README.me. (be20dbaf)

• FIX: Fetch web requests with big payloads dropping connection (#273). (425889dc)