This release adds 3 notable features for engineering teams evaluating rollout.
✓ No known CVEs patched in this version
Topics
+11 more
Affected surfaces
Summary
AI summaryUpdates feat, fix, and lucide across a mixed release.
Changes in this release
| Type | Severity | Summary | CVE |
|---|---|---|---|
| Feature | Medium |
Adds model migrations for Google integration. Adds model migrations for Google integration. Source: llm_adapter@2026-06-13 Confidence: high |
— |
| Feature | Medium |
Adds Deepseek model support. Adds Deepseek model support. Source: llm_adapter@2026-06-13 Confidence: high |
— |
| Feature | Medium |
Adds remote device connection as a tool. Adds remote device connection as a tool. Source: llm_adapter@2026-06-13 Confidence: high |
— |
| Feature | Medium |
Adds model ID field on token usage reporting. Adds model ID field on token usage reporting. Source: llm_adapter@2026-06-13 Confidence: high |
— |
| Feature | Medium |
Adds reasoning passthrough capability. Adds reasoning passthrough capability. Source: llm_adapter@2026-06-13 Confidence: high |
— |
| Feature | Medium |
Adds OpenAI Responses API support with per‑model api_flavor. Adds OpenAI Responses API support with per‑model api_flavor. Source: llm_adapter@2026-06-13 Confidence: high |
— |
| Feature | Medium |
Adds response_format / response_schema (Structured Outputs) on /v1/chat/completions endpoint. Adds response_format / response_schema (Structured Outputs) on /v1/chat/completions endpoint. Source: llm_adapter@2026-06-13 Confidence: high |
— |
| Feature | Low |
Improves mobile sheet handling. Improves mobile sheet handling. Source: granite4.1:30b@2026-06-13-audit Confidence: high |
— |
| Feature | Low |
Adds conversation visibility settings. Adds conversation visibility settings. Source: granite4.1:30b@2026-06-13-audit Confidence: high |
— |
| Feature | Low |
Enhances tool call follow‑up in compatibility mode. Enhances tool call follow‑up in compatibility mode. Source: granite4.1:30b@2026-06-13-audit Confidence: high |
— |
| Feature | Low |
Introduces OIDC login support. Introduces OIDC login support. Source: granite4.1:30b@2026-06-13-audit Confidence: high |
— |
| Feature | Low |
Improves prompt dropdown UI. Improves prompt dropdown UI. Source: granite4.1:30b@2026-06-13-audit Confidence: high |
— |
| Performance | Low |
Improves retry logic. Improves retry logic. Source: granite4.1:30b@2026-06-13-audit Confidence: high |
— |
| Performance | Low |
Improves thread handling. Improves thread handling. Source: granite4.1:30b@2026-06-13-audit Confidence: high |
— |
| Bugfix | Medium |
Fixes remote‑device tool timeout issues on scheduled runs using Redis broker. Fixes remote‑device tool timeout issues on scheduled runs using Redis broker. Source: llm_adapter@2026-06-13 Confidence: high |
— |
| Bugfix | Medium |
Fixes backup model response parsing errors. Fixes backup model response parsing errors. Source: llm_adapter@2026-06-13 Confidence: high |
— |
| Bugfix | Medium |
Fixes UI surface revocation when durable state reaches terminal status. Fixes UI surface revocation when durable state reaches terminal status. Source: llm_adapter@2026-06-13 Confidence: high |
— |
| Bugfix | Medium |
Fixes local tool calling issues on non‑OpenAI endpoints. Fixes local tool calling issues on non‑OpenAI endpoints. Source: granite4.1:30b@2026-06-13-audit Confidence: high |
— |
| Bugfix | Low |
Fixes tool icon display issues. Fixes tool icon display issues. Source: granite4.1:30b@2026-06-13-audit Confidence: high |
— |
| Refactor | Low |
Cleans up frontend components. Cleans up frontend components. Source: granite4.1:30b@2026-06-13-audit Confidence: high |
— |
| Refactor | Low |
Unifies navigation bar icons (lucide) and fixes sidebar alignment. Unifies navigation bar icons (lucide) and fixes sidebar alignment. Source: granite4.1:30b@2026-06-13-audit Confidence: high |
— |
| Refactor | Low |
Revamps default prompt presets, templating, and tool naming. Revamps default prompt presets, templating, and tool naming. Source: granite4.1:30b@2026-06-13-audit Confidence: high |
— |
| Refactor | Low |
Revamps analytics and logs with per‑agent attribution and unified timeline. Revamps analytics and logs with per‑agent attribution and unified timeline. Source: granite4.1:30b@2026-06-13-audit Confidence: high |
— |
Full changelog
What's Changed
- feat: model migrations google by @dartpain in https://github.com/arc53/DocsGPT/pull/2496
- Frontend component cleanup by @dartpain in https://github.com/arc53/DocsGPT/pull/2495
- feat: add deepseek by @dartpain in https://github.com/arc53/DocsGPT/pull/2497
- Deepseek by @dartpain in https://github.com/arc53/DocsGPT/pull/2499
- Sheet mobile by @dartpain in https://github.com/arc53/DocsGPT/pull/2502
- feat: remote device connection as tool by @dartpain in https://github.com/arc53/DocsGPT/pull/2506
- feat: add model id on token usage by @dartpain in https://github.com/arc53/DocsGPT/pull/2507
- Fix tool icons by @dartpain in https://github.com/arc53/DocsGPT/pull/2510
- Fix remote-device tool timing out on scheduled runs (Redis-backed broker) by @dartpain in https://github.com/arc53/DocsGPT/pull/2511
- Unify nav bar icons (lucide) and fix sidebar alignment by @pabik in https://github.com/arc53/DocsGPT/pull/2513
- Better retry by @dartpain in https://github.com/arc53/DocsGPT/pull/2517
- fix: backup model response parsing by @dartpain in https://github.com/arc53/DocsGPT/pull/2518
- feat: resoning passthrough by @dartpain in https://github.com/arc53/DocsGPT/pull/2520
- feat: OpenAI Responses API support (per-model api_flavor) by @dartpain in https://github.com/arc53/DocsGPT/pull/2521
- fix: issues with local tool calling on non openai endpoint by @dartpain in https://github.com/arc53/DocsGPT/pull/2522
- feat(v1): response_format / response_schema (Structured Outputs) on /v1/chat/completions by @dartpain in https://github.com/arc53/DocsGPT/pull/2523
- fix: revoke stale UI surfaces when durable state goes terminal by @dartpain in https://github.com/arc53/DocsGPT/pull/2526
- Better thread by @dartpain in https://github.com/arc53/DocsGPT/pull/2527
- feat: conv visibility by @dartpain in https://github.com/arc53/DocsGPT/pull/2528
- feat: better tool call followup on compat mode by @dartpain in https://github.com/arc53/DocsGPT/pull/2529
- feat: visibility settings by @dartpain in https://github.com/arc53/DocsGPT/pull/2531
- feat: oidc login by @dartpain in https://github.com/arc53/DocsGPT/pull/2530
- Revamp default prompt presets, prompt templating, and tool naming by @dartpain in https://github.com/arc53/DocsGPT/pull/2533
- feat: better prompt dropdown by @dartpain in https://github.com/arc53/DocsGPT/pull/2536
- revamp analytics & logs — per-agent attribution, unified log timeline by @pabik in https://github.com/arc53/DocsGPT/pull/2534
Full Changelog: https://github.com/arc53/DocsGPT/compare/0.17.2...0.17.3
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About DocsGPT
Private AI platform for agents, assistants and enterprise search. Built-in Agent Builder, Deep research, Document analysis, Multi-model support, and API connectivity for agents.
Related context
Related tools
Beta — feedback welcome: [email protected]