| v0.5.91 | ## Highlights - Temporary provider-capacity refusals no longer trip the hourly automated-start circuit. Fleet Manager keeps dispatching through short OpenAI capacity pressure, while each fresh delegated or scheduled start still observes the provider's one-minute retry floor before trying again ([#4985](https://github.com/tale-project/tale/pull/4985), [#4945](https://github.com/tale-project/tale/issues/4945)). Real automated starts continue to count toward the circuit, so repeated task failures | High | 10/15/2026 |
| v0.5.90 | ## Highlights - [`tale deploy` upgrades a workspace from v0.5.77 across the automation writer protocol cutover](https://github.com/tale-project/tale/pull/4954). From v0.5.79 to v0.5.89 it refused every such workspace with "An existing protocol-1 database must upgrade through managed deployment apply…", and managed apply cannot take a `tale deploy` workspace, so these workspaces could not leave v0.5.77. Before it changes anything, `tale deploy` now checks the running backend's release, the datab | High | 10/14/2026 |
| v0.5.89 | ## Highlights - Native deployment provisioning now accepts the Better Auth 1.7 built-in password capability. Upgrading a host that already runs v0.5.87 or v0.5.88 completes the break-glass administrator reconciliation instead of stopping after the service containers become healthy. ## Upgrade notes - From v0.5.88 to v0.5.89 no database migration or manual configuration change is needed. The release is a deployment hotfix for the native CLI; rerun the normal Tale CLI deployment/adoption flow a | High | 10/13/2026 |
| v0.5.86 | ## Highlights - The CLI release lane now builds from the same source-bound workflow contract as service images, so native binaries and their release tag cannot silently come from different commits. - Release metadata and image identity checks now exercise the exact prepared source SHA used by the publication job. ## Upgrade notes - Adopt v0.5.86 as one source-bound cohort through the normal Ops release workflow. It supersedes v0.5.85 for the managed deployment CLI and runtime; no database mig | High | 10/11/2026 |
| v0.5.83 | ## Highlights - Standing roles now wake when a project worker finishes, so a released slot can be refilled by the fleet manager without waiting for the next scheduled occurrence. Wake claims are scoped to the project and preserve the native task and review guards. [#4608](https://github.com/tale-project/tale/pull/4608) - Audit-log sealing runs asynchronously after the write path, while the integrity panel and verification checks continue to expose the exact source and chain state. The per-proje | High | 10/11/2026 |
| v0.5.79 | ## Highlights - API-key creation offers a custom expiry option and previews the resulting expiry date. The keys table now includes an **Expires** column. [#4495](https://github.com/tale-project/tale/pull/4495) - Owners and Admins can create an organization-bound API key for another eligible member, a team, a project or the organization itself. A member's key follows that member's live permissions; team, project and organization keys act with their own assigned role and keep working wh | High | 10/10/2026 |
| v0.5.77 | ## Highlights - When an agent's workspace has reached its limit of four concurrent executions, a new run waits for a place instead of failing. Finishing a run wakes the oldest waiter in that workspace; waiting does not consume an automatic retry. [Task automation](https://docs.tale.dev/platform/projects/task-automation) · [#4481](https://github.com/tale-project/tale/pull/4481) - Coding agents can discover their granted workspace tools even when they call `workspace_status` through the | High | 10/7/2026 |
| v0.5.75 | ## Highlights - First pages load faster. Each page now brings its interface text by topic, together with the code that reads it, instead of the whole catalog: the cold load drops from 891 to 792 KB gzip, and every first page carries 26 KB gzip less render-blocking CSS. A German or French session fetches about 36 KB gzip of its language before the first frame instead of the whole 124 KB catalog. Switching language fetches the new language's text for the pages already open, and the next | High | 10/7/2026 |
| v0.5.74 | ## Highlights - Connected tasks can follow the source's full business workflow. A source integration can publish actions with required fields for verification, closure or reopening, including two stages that share the same Tale column. People submit the complete form from the task's details; the source checks their verified identity, permissions and business rules before accepting the change. Pending requests and accepted or refused decisions remain visible after a reload. The new API | High | 10/6/2026 |
| v0.5.72 | ## Highlights - Chat can use OpenAI's GPT-6 models. GPT-6 Astra, GPT-6 Sol, GPT-6 Luna and GPT-6.1 Sol are offered through an OpenAI API key or environment-variable credential, and Tale calls Astra and GPT-6.1 Sol through the Responses API, the only way they make tool calls. Project agents and automations run those two on Codex; another runtime now says so instead of refusing without a reason. Subscription credentials are marked **Tasks and automations only** in Settings, and the chat | High | 10/4/2026 |
| v0.5.71 | ## Highlights - Codex agents can select GPT-6.1 Sol through an OpenAI subscription credential. The model catalog now records its Responses API requirement, so incompatible native chat and agent runtimes refuse it before sending a tool request. [Provider setup](https://docs.tale.dev/platform/admin/providers) - Subscription broker cooldowns immediately following the same agent's rate-limit failure no longer consume an extra automated-start allowance. Manager agents can read whether a fa | High | 10/3/2026 |
| v0.5.69 | ## API contract changes None in this range. The contract stays at 3.9.0: 139 operations. ## What's Changed * feat(platform): filter a website's pages to the failed or skipped ones by @larryro in https://github.com/tale-project/tale/pull/4065 * fix(platform): recheck task access before notification delivery by @yannickmonney in https://github.com/tale-project/tale/pull/4056 **Full Changelog**: https://github.com/tale-project/tale/compare/v0.5.68...v0.5.69 | High | 10/2/2026 |
| v0.5.68 | ## API contract changes The contract moved from 3.8.0 to 3.9.0 (139 → 139 operations). Read the API reference's versioning section before upgrading a pinned client. Added operations: none. Removed operations: none. ### Changelog 3.9.0 — 2026-10-01: project agents answer `managed` — true for the organization's standard agent, which Tale sets up in a project without agents of its own under the new `standard_agent` governance policy and keeps in line with it. Saving one answers 409 `PROJECT_AG | High | 10/2/2026 |
| v0.5.67 | 4 PRs since v0.5.66. Two migrations (0142, 0143), both rolling-deploy safe, applied at boot. No proxy, compose or `.env` change: `tale update`, then `tale deploy` as usual. ## Projects and tasks - A run that fails for good is said and told: the task shows **The agent couldn't finish this task** with the cause, who can fix it, **Details** and **Retry**; the run's starter and the task's watchers get the new **Agent run failed** notification (email when a mailbox is connected), under the existing | High | 10/1/2026 |
| v0.5.64 | ## Publication incomplete The image Release workflow passed, but all five tag-bound CLI builds failed synthetic tag-validation tests; this release has no CLI assets and is not ready for the Ops deployment pipeline. The cause is tracked in [#3970](https://github.com/tale-project/tale/issues/3970), with the [failed CLI run](https://github.com/tale-project/tale/actions/runs/36658878748) retained. The `v0.5.64` tag remains fixed at `ebf4546fb1455a236c1046f3d73ef78c2e1d0109`. A corrected, fully vali | High | 9/30/2026 |
| v0.5.63 | ## API contract changes The contract moved from 3.3.0 to 3.5.0 (136 → 136 operations). Read the API reference's versioning section before upgrading a pinned client. Added operations: none. Removed operations: none. ### Changelog 3.5.0 — 2026-09-29: an organization may reserve organization-wide skills. Under its `skill_sharing` policy (Editors and above, or owners and administrators, plus members granted the `tale:skills.publish` capability), `PUT /api/v1/skills/{slug}` answers 403 `SKILL_PU | High | 9/29/2026 |
| v0.5.61 | One merged pull request (#3503) and nineteen direct commits since v0.5.60. No migrations, no API contract change (2.1.0), and no proxy or sandbox-runtime image change: run `tale update` and then `tale deploy` as usual. ## Chat - A project member opening a chat someone else shared with the project now reads its messages under the read-only pill, and sees the version the owner has selected rather than the root's replaced branch; Copy, message info and Export stay, nothing that writes. (#3503) ## | High | 9/27/2026 |
| v0.5.60 | One merged pull request (#3498, the S3 fixes from the 2026-09-26 platform evaluation) and fourteen direct commits since v0.5.59. Migrations `0119`–`0121` (backfills, applied at boot); API contract 2.0.0 → 2.1.0; the proxy image changes, so run `tale update` and then `tale deploy --stop` (recreates db, object-store and proxy with a brief downtime). ## Behaviour changes - Mounting the WebDAV root `/dav/<org>/` now lists two folders, `documents/` and `.trash/`; dot-segment paths (`..`, `%2e%2e`) a | High | 9/27/2026 |
| v0.5.55 | ## What's Changed * fix(ai-gateway): keep the account list when a re-read fails by @yannickmonney in https://github.com/tale-project/tale/pull/3487 * fix(platform): put the email and profile claims back into Tale's ID tokens by @yannickmonney in https://github.com/tale-project/tale/pull/3488 **Full Changelog**: https://github.com/tale-project/tale/compare/v0.5.54...v0.5.55 | High | 9/25/2026 |
| v0.5.51 | Ten merged pull requests since v0.5.50 (#3466–#3475). One migration (0113, additive, applied on start); `tale deploy` as usual. ## Inbox - A reply leaves from the mailbox that received the thread; an org with two mailboxes on one email connector no longer answers from the default one. (#3468) - Rows show the channel a conversation arrived on; the header names the connector or API source; the composer says where the reply goes; the Channel filter works again. (#3470) - Picking the current assign | High | 9/24/2026 |
| v0.5.50 | Six merged pull requests since v0.5.49 (#3460–#3465). No migration; `tale deploy` as usual. ## Inbox - One **Filter** button holds every facet: assignee (people and teams), read status, source. (#3464) - The Inbox nav item and the mobile tab bar show the unread count. (#3461) - Composing to an address the org does not know offers **Add "…" as a contact**; the recipient picker searches server-side. (#3465) ## Chat - Chat follows the custom instructions saved under Settings › Preferences. The Me | High | 9/22/2026 |
| v0.5.49 | ## What's Changed * fix(ui): never restore focus to the menu item a dialog was opened from by @yannickmonney in https://github.com/tale-project/tale/pull/3459 * fix(ai-gateway): dress the panel in the design system it builds on by @yannickmonney in https://github.com/tale-project/tale/pull/3458 **Full Changelog**: https://github.com/tale-project/tale/compare/v0.5.48...v0.5.49 | High | 9/22/2026 |
| v0.5.45 | **0.5.45 carries four merged pull requests and one direct commit — every change since 0.5.44.** Three numbers have gone by since the last curated notes, 0.5.41. `v0.5.42` was tagged on 2026-09-20 but its container test gate failed — the pull request it was cut on had just switched the gate from rebuilding every image on the runner to validating the pulled release images, and the sandbox spawner's local image aliases were not yet re-tagged from those pulls, which the next number added — so no mul | High | 9/21/2026 |
| v0.5.39 | **0.5.39 carries 1 merged pull request.** It closes every finding of the 2026-09-19 external evaluation of the machine surface (the eleventh round, run against 0.5.36 and contract 1.17.0), the first round that found nothing above the third severity: seven S3 and twenty-seven S4, each re-confirmed in code before it was touched — one, a pagination marker the evaluator read without resolving a reference, needed nothing. The largest fixes are the ones a client feels: an organization's teams can be l | High | 9/20/2026 |
| v0.5.36 | **0.5.36 carries 1 merged pull request.** A team becomes what it always claimed to be: a label on work that says who may see it, and a queue a conversation can wait in — never a workspace to switch into. A document, folder or project carries the teams that may see it; empty means the whole organization, and Owners and Admins see everything. One rule states this once and every door reads it, including the doors that ignored teams until now — WebDAV, folder creation, the cloud imports, the documen | High | 9/19/2026 |
| v0.5.33 | **0.5.33 carries 5 merged pull requests.** Three of them are about who may hold an account on a deployment and what that account is worth: the account-creation route now closes the moment a deployment holds an account, an account an administrator provisions counts as a verified one everywhere at once, and a managed deployment can rename its deploy operator in place and declare a break-glass administrator whose password never reaches the deployment. The fourth closes the 2026-09-18 external API e | High | 9/18/2026 |
| v0.5.31 | **0.5.31 carries 17 merged pull requests.** Six of them do one thing: a managed agent turn against a model you host yourself now finishes. Six changes close the separate ways such a turn used to be re-sent, abandoned, un-cached or silently settled as "done". Beside that, chat gains an organization-wide **Audio transcription model** setting and stops warning about audio nobody asked it to transcribe, a UI evaluation round's findings are fixed across dialogs, settings and governance, two paginated | High | 9/17/2026 |
| v0.5.30 | **0.5.30 is a two-change patch on the 0.5 line**, and both changes are for a deployment an operator assembled themselves: a healthy stack no longer reports **Service outage** on its public status page, and the Own Compose guide finally lists the egress proxy's real capability set and the IPv6 sysctls it has needed since 0.5.18. Nothing changes shape — no migration, no contract change, no configuration file, no image in the stop-gated tier. The upgrade is `tale update` followed by a plain `tale d | High | 9/16/2026 |
| v0.5.28 | **0.5.28 is a small patch on the 0.5 line.** Two changes since 0.5.27: a broken OneDrive or Google Drive folder sync is now visible in Documents and its owner is told once, and the ninth external API evaluation's findings are closed at their root. It ships one forward-only migration on the application database, an API contract that moves from 1.11.0 to 1.12.0 — additive, plus one correction that makes the document valid to a strict OpenAPI 3.0 client — and nothing else that changes shape: no env | High | 9/15/2026 |
| v0.5.25 | **0.5.25 is a fix release on the 0.5 line, and this one stays inside the patch promise: no schema migration, no configuration-file change, no new environment variable, and the API contract stays at 1.6.0.** It closes the sixth external API evaluation's findings, every one a gap between what the surface does and what it said rather than a wire defect: the OpenAPI document and the reference now say what they were silent on, three refusals carry what a client can branch on, `Strict-Transport-Securi | High | 9/14/2026 |
| v0.5.23 | **0.5.23 is a fix release on the 0.5 line that carries more than its fixes: the fourth set of REST contract corrections from an external API evaluation (contract 1.4.0 → 1.5.0), a repaired and now enforced rule for caller-owned external keys, observable crawl failures, an honest 408 at the door, and a server-pinned scope for the chat assistant inside a project.** Two platform migrations (0098, 0099) and one knowledge-database migration apply themselves on `tale deploy`, each written to run while | High | 9/13/2026 |
| v0.5.21 | **0.5.21 is a fix release on the 0.5 line with one fix, in the CLI: managed deployments complete again.** Every `tale deploy --bundle` since 0.5.20's CLI started keeping its backend-local state under the backend's data volume ended after the runtime had already rolled, with the backend-local provisioning phase refused by its own ownership check. No migration, no config-file change, no platform image change beyond the version stamp. The Known issues from 0.5.20 are unchanged. ## Highlights ### | High | 9/12/2026 |
| v0.4.18 | ## What's Changed * feat(web): show how we work with companies on the about page by @yannickmonney in https://github.com/tale-project/tale/pull/3080 * fix(platform): show review-waiting count on task cards by @Israeltheminer in https://github.com/tale-project/tale/pull/3085 * feat(platform): add glm-5.3-flash to the zai model catalog by @larryro in https://github.com/tale-project/tale/pull/3099 **Full Changelog**: https://github.com/tale-project/tale/compare/v0.4.17...v0.4.18 | High | 8/28/2026 |
| v0.4.9 | **0.4.9 is a fix release on the 0.4 line — an in-place update: `tale update`, then `tale deploy`.** No data migrations ship with it. ## Highlights **Chat `web_fetch` pages like `rag_fetch` — a long page is no longer a dead end.** The tool used to clip to the first 20,000 characters with no way to continue, so a longer page (a 39.7k-char weather JSON, a long docs article) sent the model into refetch loops until the turn's step limit. Optional `offset`/`limit` now select a window of up to 20,000 | High | 8/17/2026 |
| v0.4.0 | **0.4.0 is a breaking minor: it requires a fresh deployment.** There is no upgrade path between 0.3.x and 0.4 — in either direction. If you operate a 0.3.x instance, read the breaking changes below before touching anything. ## Highlights 0.4 rebuilds Tale's AI backend, and with it the data model, from a clean baseline on a minimal YAML-config core (#2857). Automations are versioned configurations you upload and deploy in one click; task agents work a task end to end, with a live step timeline, | High | 8/4/2026 |
| v0.3.12 | ## Upgrade behaviour change: `tale update` now stays on your release line `tale update` (without `--version`) now targets the newest release **within your current x.y release line** — a 0.3.x instance updates to the newest 0.3.x and never crosses to a newer line on its own. When a newer line exists, the command says so and prints the exact `tale update --version <ver>` command to move deliberately. Explicit `--version` behaves as before (any version, up or down). Why this matters now: the upco | Medium | 7/24/2026 |
| v0.3.9 | ## What's Changed * perf(platform): two-tier chat tool gating — <10k-token trivial turns by @yannickmonney in https://github.com/tale-project/tale/pull/2782 * feat(platform): probe the node-action lane on the status page by @larryro in https://github.com/tale-project/tale/pull/2795 * feat(docs): video-pipeline authoring loop — check, doctor, mock rehearsal, draft compose, A/V verification by @yannickmonney in https://github.com/tale-project/tale/pull/2796 * feat(platform): unified collapsible ap | High | 7/19/2026 |
| v0.3.3 | ## What's Changed * fix(platform): polish catalogs, settings actions, and governance banners by @AdeolaAdekoya in https://github.com/tale-project/tale/pull/2515 * fix(platform): skip [1m] suffix on gateway model refs for external agents (#2396) by @larryro in https://github.com/tale-project/tale/pull/2505 * fix: agent bound skills tab is not scrollable (#2487) by @larryro in https://github.com/tale-project/tale/pull/2501 * fix(platform): add spacing between default model label and sandbox badge | Medium | 7/7/2026 |
| v0.2.97 | ## What's Changed * fix(sandbox): stage session exec temp on the workspace, not /tmp by @larryro in https://github.com/tale-project/tale/pull/2321 * docs(platform): correct SCIM delete wording to match behaviour by @larryro in https://github.com/tale-project/tale/pull/2323 * feat(platform): unambiguous multi-org SSO sign-in with org picker by @larryro in https://github.com/tale-project/tale/pull/2322 **Full Changelog**: https://github.com/tale-project/tale/compare/v0.2.96...v0.2.97 | High | 7/2/2026 |
| v0.2.86 | ## Tale v0.2.86 ### Website crawling now runs in-process (#1935) Registered websites are now re-crawled by an in-process Convex cron (`scan due websites`, every 5 minutes) — the replacement for the former standalone crawler service's poll loop. Without it, registered websites never re-crawled after the standalone service was removed. Full-site sitemap discovery is also restored (the interim 100-URLs-per-scan cap was a regression); the discovery ceiling is tunable via the optional `CRAWLER_MAX_D | High | 6/23/2026 |
| v0.2.84 | # v0.2.84 — Kubernetes sandbox backend, adaptive reasoning, and audit hardening This release adds a Kubernetes execution backend for the sandbox alongside the existing Docker path, a major LLM-pipeline upgrade (adaptive reasoning budgets, prompt caching, and a redesigned Auto router), and new knowledge-base workflows in chat. It also hardens audit-log access, clears the transitive HIGH dependency advisory backlog, and makes `tale rollback` safe by design. ## 🔒 Security - Audit-log reads are | High | 6/11/2026 |
| v0.2.83 | # v0.2.83 — Reasoning latency metrics and a leaner chat path This release adds reasoning-aware time-to-first-token metrics, so you can see what you actually wait for on a reasoning model — the "thinking" stream that starts before the first answer token — plus a send-relative breakdown that exposes the backend "setup before model" overhead the old action-relative metric missed. It then cuts that overhead with a consolidated governance round-trip and pre-stream parallelization, enforces role-scop | High | 6/8/2026 |
| v0.2.81 | # v0.2.81 — Tasks, collaboration, and session security This release adds a task platform with a collaborative inbox and project-scoped secrets, plus Slack and Confluence integrations and a generic OIDC SSO adapter. It hardens sessions with optional server-side idle timeout and adds daily audit-log integrity verification, alongside dependency CVE patches and a large batch of UI and import fixes. ## 🔒 Security - Server-side session idle timeout: set the optional `SESSION_IDLE_TIMEOUT_MINUTES` | High | 6/2/2026 |
| v0.2.76 | # v0.2.76 — Skill upload hardening Closes data-safety and UX gaps in the skills upload flow surfaced by multi-agent review. Adds cross-tenant storage isolation, per-slug exclusion locking, blob lifecycle cleanup, and protects UI-uploaded skills from being clobbered by `tale deploy`. No schema-breaking changes; no manual migration required. ## 🔒 Security - Bind each `_storage` blob to the uploading org via a new `skillUploadIntents` table; `uploadSkillBundle` now refuses any `storageId` not o | High | 5/26/2026 |
| v0.2.72 | # v0.2.72 — Confidentiality notice refresh and example library cleanup A small maintenance release. The chat confidentiality footer gets clearer copy and a centered layout across en/de/fr, and the shipped example-agent library is trimmed to the actively maintained set. ## 🛠 Improvements - Refresh the chat confidentiality footer: shorter copy that emphasizes verifying AI responses and avoiding sensitive data, with the icon and text now centered horizontally. Strings updated in en, de, and fr. | High | 5/12/2026 |
| v0.2.69 | # v0.2.69 — Marketing site, optional SOPS, version-pinned installs This release ships the public marketing site at `services/web` (with a shared `@tale/ui` library and `plop` scaffolding), makes SOPS encryption optional for provider secrets, and adds version-pinned CLI installs and downgrades. It also closes a thread-access regression on deleted-org and ex-member states, and stops the failover loop when no provider is configured. ## 🔒 Security - Closed access bypass and Convex subscription l | High | 5/6/2026 |
| v0.2.67 | # v0.2.67 — Save agents before adding API keys Fixes a save-time validation bug that blocked agents from referencing providers whose JSON config existed but didn't yet have an API key. Provider configuration and secret provisioning can now happen in any order, while runtime invocation continues to enforce key availability. ## 🐛 Fixes - Allow saving agents that reference keyless providers — `saveAgent` now validates `supportedModels` against the configured-models list (read directly from provi | High | 4/30/2026 |
| v0.2.63 | # v0.2.63 — i18n cleanup, docs polish, and DB startup fix Unblocks self-hosted startup when the database init scripts are still running, lowers the default RAG similarity threshold to improve agent knowledge recall, and lands a large translations + documentation pass with new pages on 2FA, image generation, automation metrics, and chat basics across all three locales. ## 🔒 Security - **`brace-expansion` override pinned to `4.0.1`.** The repo-wide `overrides` entry now forces every transitive | High | 4/27/2026 |
| v0.2.54 | # v0.2.54 — RAG recall and UI polish Patch release tuning RAG retrieval defaults and fixing two platform UI issues: PDF preview horizontal overflow and 2FA policy settings not persisting toggle/grace-period edits. ## 🐛 Fixes - Lower default RAG similarity threshold from 0.51 to 0.4 so semantically relevant chunks aren't filtered out before ranking; debug logs now include the top vector score to make threshold tuning observable (#1598) - Fix horizontal scroll on the document PDF preview so wid | High | 4/22/2026 |
| v0.2.53 | # v0.2.53 — Content safety guardrails Ships an end-to-end content-governance pipeline for chat: admin-configurable keyword/regex filters, built-in PII detection, and an external moderation-provider adapter with presets for OpenAI, Azure Content Safety, and Perspective. All three run on both user input and streamed agent output, with mask / block / flag outcomes wired into a new Recent Events feed. ## 🚀 Features - Chat filter, PII, and moderation-provider guardrails with per-stage mask/block/f | High | 4/21/2026 |
| v0.2.52 | # v0.2.52 — Composer capability pills & smarter language handling Active composer capabilities now surface as dismissible pills on the chat toolbar, and picking an agent that needs an un-installed integration deep-links straight to its setup panel. The default chat agent prompt was also reworked with a strict language-priority rule and user-profile context. ## 🤖 Model & Provider - Default "Assistant" chat agent prompt: language handling rewritten as a strict priority order (explicit request → | High | 4/21/2026 |
| v0.2.51 | # v0.2.51 — Audio transcription for chat uploads Chat uploads now accept audio files: drop in an MP3/WAV/M4A and it's transcribed via OpenAI Whisper and indexed into RAG so chats can cite it. Plus the chat history toggle gets a clock icon, the save-prompt menu splits out, and a few prompt and default tweaks. ## 🚀 Features - **Audio transcription for chat uploads** (#1591): audio attachments are auto-transcribed via OpenAI Whisper (`whisper-1`) and indexed into RAG. Adds a `transcription` mode | High | 4/21/2026 |
| v0.2.50 | # v0.2.50 — Vercel AI Gateway seed provider Ships a Vercel AI Gateway seed provider config for image generation (FLUX, Imagen, Nano Banana), and simplifies the `.gitignore` so all `examples/providers/*` seed files are tracked by default rather than allow-listed one-by-one. ## 📝 Other - Track all `examples/providers/*` seed files and add `vercel-gateway.json` + `vercel-gateway.secrets.json` with FLUX 2 Pro, FLUX 1.1 Pro Ultra, FLUX.1 Kontext (Pro/Max), Imagen 4 (Standard/Fast/Ultra), and Gemin | High | 4/20/2026 |
| v0.2.49 | # v0.2.49 — Governance default × model access alignment Fixes an opaque "Failed to send message" error that users hit when an admin's governance default model was excluded by the organization's `model_access` allowlist or blocklist. Closes related gaps where the OpenAI-compat endpoint and workflow LLM steps were only enforcing budget — they now enforce `model_access` too. Adds clearer error surfaces and admin-side conflict warnings. ## 🔒 Security - Enforce `model_access` on `openai_compat` `c | High | 4/20/2026 |
| v0.2.48 | # v0.2.48 — Multi-tenancy, TOTP, and a documented Tale This release is mostly additive: TOTP two-factor auth, per-org provider isolation, workflow & usage analytics dashboards, agent todos with a live plan pane, FLUX/Gemini image agents, an OpenAI-compatible agent webhook (Meetily), and a central retention & deletion policy. CI is hardened with pinned action SHAs, scheduled Trivy/bun-audit scans, and merge-queue support. Docs are restructured by persona, internationalized across en/de/fr, and g | High | 4/20/2026 |
| v0.2.47 | ## What's Changed ### 🐛 Bug Fixes - Disable SSE file events by default and fix stream cancel bug (#1560) ### 📝 Other Changes - Fix empty catch blocks in SSE-related code ## Upgrade Run `tale upgrade` to update the CLI, then `tale deploy` to apply the new version. No manual migration required. ## Contributors @larryro --- **Full Changelog**: https://github.com/tale-project/tale/compare/v0.2.46...v0.2.47 | High | 4/17/2026 |
| v0.2.46 | ## What's Changed ### 🚀 Features - Provider model fetching with search and batch selection (#1557) - Add member count and created date to teams table (#1556) - Sticky save bars and conditional save buttons in settings (#1555) - Add French translations and sync all locale files ### 🐛 Bug Fixes - Allow microphone access in permissions policy (#1559) ### 📝 Other Changes - Fix missing trailing newlines in locale files - Add dev setup docs and gitignore provider files (#1558) - Extract shared t | High | 4/17/2026 |
| v0.2.45 | ## What's Changed ### 🐛 Bug Fixes - Prevent feedback comment box from pushing the fork button out of view (#1554) - Show user emails in usage table and add skeleton loading to governance tabs (#1553) - CLI upgrade now falls back to the latest release with a ready binary instead of failing when the newest release has no binary uploaded yet (#1552) ## Upgrade No manual migration required. Run `tale upgrade` then `tale deploy` to update. ## Contributors @larryro --- **Full Changelog**: https: | High | 4/16/2026 |
| v0.2.44 | ## What's Changed ### 🚀 Features - **Improved web tool with explicit fetch/search modes** — The web tool now uses a discriminated union (`mode: "fetch"` / `mode: "search"`) instead of heuristic URL detection. Fetch mode works with any public URL; search mode explicitly targets the organization's knowledge base. The tool description now clearly communicates what each mode does and when to use it. (#1551) - **Citation header for web fetch results** — Fetch mode now returns a structured citation | High | 4/16/2026 |
| v0.2.43 | ## What's Changed ### 🚀 Features - **Configurable password policy with rotation** — Admins can now define password complexity rules (min length, character requirements) and optional rotation periods via the governance settings. Passwords that exceed the rotation window trigger a forced-change screen on next login. A new `userPasswordMetadata` table tracks per-user credential timestamps independently of OAuth token refreshes. Admin-created members are prompted to change their password on first | High | 4/16/2026 |
| v0.2.42 | ## What's Changed ### 🚀 Features - **Login rate limiting** with exponential backoff and account lockout to defend against brute-force attempts (#1542) - **Centralized security headers** in Hono with a hardened Content Security Policy (#1541) - **Local hosting of libraries, fonts, and assets** for GDPR compliance and offline operation (#1543) - **Notifications panel** with mark-all-read, filtering, and pagination (#1546) - **Tabs pattern applied to all governance settings groups** for consisten | High | 4/15/2026 |
| v0.2.41 | ## What's Changed ### 🐛 Bug Fixes - Reorder fork button after feedback and space message actions (#1538) - Eliminate arena mode white flash during thread creation (#1537) ### 📝 Other Changes - Update actions/github-script action to v8 (#1494) ## Upgrade Run `tale upgrade` to update the CLI, then `tale deploy` to apply the new version. No manual migration required. ## Contributors @larryro, @renovate[bot] --- **Full Changelog**: https://github.com/tale-project/tale/compare/v0.2.40...v0.2 | High | 4/15/2026 |
| v0.2.40 | ## What's Changed ### 🐛 Bug Fixes - Add proper arena exit cleanup with branch merging — when exiting arena mode, Thread B is now properly cleaned up and its messages are merged into Thread A when B wins the verdict (#1535) - Remove duplicate Share item from chat header dropdown (#1534) ## Upgrade No manual migration required. Run `tale deploy` to update. ## Contributors @larryro --- **Full Changelog**: https://github.com/tale-project/tale/compare/v0.2.39...v0.2.40 | Medium | 4/14/2026 |
| v0.2.39 | ## What's Changed ### 🐛 Bug Fixes - **Eliminate chat loading delays with early generation status** — Moves the `generationStatus` update to the very start of the chat action (before PII, config, and budget checks), so the Convex subscription delivers `isGenerating=true` to the client with minimal delay. Pre-allocates the stream ID in the same early mutation and reuses it downstream, removing redundant stream creation. (#1533) - **Show optimistic messages instantly in standard mode** — Standar | Medium | 4/14/2026 |
| v0.2.38 | ## What's Changed ### Bug Fixes - **Arena mode**: Eliminate delays in arena mode for instant optimistic UX — "Thinking" indicator now appears immediately on send, and split view renders without flashing the single-chat view (#1531) - **OAuth2 integrations**: Automatically install integration before saving OAuth2 credentials — fixes failure when configuring OAuth2 on an uninstalled integration (#1532) ## Upgrade No manual migration required. Run `tale deploy` to update. --- **Full Changelog** | Medium | 4/14/2026 |
| v0.2.37 | ## What's Changed ### Bug Fixes - **Web search**: Add vector similarity pre-filter to web search — queries that are semantically irrelevant now return empty results instead of noisy keyword matches (#1530) - **Search quality**: Raise default similarity threshold from 0.3 to 0.4 across RAG, web context, and crawler search pipelines for more relevant results ## Upgrade No manual migration required. Run `tale deploy` to update. --- **Full Changelog**: https://github.com/tale-project/tale/compar | Medium | 4/14/2026 |
| v0.2.36 | ## What's Changed ### 🐛 Bug Fixes - Add optimistic message display for arena mode — messages now appear immediately in the chat while waiting for the server response (#1529) - Fix string metadata parsing, deduplicate citation indices, and add locale description to organization settings (#1528) - Always run migration `detect()` to catch end-state drift — previously, recorded migrations could be skipped even when their target state had drifted (#1527) ## Upgrade No manual migration required. R | Medium | 4/14/2026 |
| v0.2.35 | ## What's Changed ### 🐛 Bug Fixes - **cli:** Ensure Docker network exists before stateful deploy, move Convex to stateful compose, add Convex to CI (#1526) - Convex is now managed as a stateful service (alongside db and proxy) instead of being emitted in the blue/green color compose — fixes lifecycle issues where Convex could be incorrectly torn down during blue/green rotations - Docker network is now created before any `docker compose up`, preventing startup failures after migrations that | Medium | 4/14/2026 |
| v0.2.34 | ## What's Changed This release splits the Convex backend into its own Docker service. The data migration from `platform-data` to a new `convex-data` volume runs automatically on the next `tale start` / `tale deploy` after upgrading (gated behind a confirmation prompt). ### 🚀 Features - Split Convex backend into its own service (#1511) - New `convex` Docker service with its own `convex-data` volume - Platform container slimmed significantly (Convex backend, Dashboard, and seed assets moved | Medium | 4/13/2026 |
| v0.2.32 | ## What's Changed ### 🚀 Features - Add language switcher and sync i18n with browser locale (e5c15ad) ### 🐛 Bug Fixes - Use public download URL for audit log export (#1495) - Use correct provider credentials and support local dev config sync (#1476) - Rename team filter "All teams" to "All" for clarity (7844697) ### 📝 Other Changes - Overhaul settings UI with grouped layouts and enhanced panels (#1485) ## Upgrade Run `tale deploy` to update to this version. No manual migration required. | Medium | 4/13/2026 |
| v0.2.33 | ## What's Changed ### 🚀 Features - **CLI: Per-project Docker resource isolation** (#1497) — Each Tale project now has a unique id (stored in `tale.json`) that namespaces all Docker volumes, containers, and networks, so multiple Tale projects can coexist on one machine without colliding. - **CLI: Host service access from containers** (#1497) — Containers can now reach host services (e.g. Ollama on `localhost:11434`) via `host.docker.internal`. Applied to `platform`, `rag`, `crawler`, and `proxy | Medium | 4/13/2026 |
| v0.2.31 | ## What's Changed ### 🐛 Bug Fixes - **Enforce budget max request limits for all period types** (#1464) — Budget enforcement now tracks usage across daily, weekly, and monthly periods simultaneously instead of only the configured period. The usage ledger writes one row per period per team, and queries aggregate all matching rows. The budget banner was extracted into its own component and now respects the selected team filter. Switching teams in the user menu navigates back to chat so the banner | Medium | 4/12/2026 |
| v0.2.30 | ## What's Changed ### 🚀 Features - Add editing support to Canvas with apply-back to message (#1451) ### 🐛 Bug Fixes - Use forkedAt timestamp for fork divider and fix order source (#1453) ## Upgrade Run `tale deploy` to update to this version. No manual migration required. ## Contributors @larryro --- **Full Changelog**: https://github.com/tale-project/tale/compare/v0.2.29...v0.2.30 | Medium | 4/12/2026 |
| v0.2.29 | ## What's Changed ### 🐛 Bug Fixes - Preserve first data row in manual CSV import (#1449) ## Upgrade Run `tale deploy` to update to this version. No manual migration required. ## Contributors @larryro --- **Full Changelog**: https://github.com/tale-project/tale/compare/v0.2.28...v0.2.29 | Medium | 4/12/2026 |
| v0.2.28 | ## What's Changed ### 🐛 Bug Fixes - **RAG: Apply similarity threshold as vector pre-filter before RRF merge** — The similarity threshold is now applied to vector search results *before* the Reciprocal Rank Fusion merge instead of as a post-filter. When all vector results fall below the threshold, FTS-only keyword noise is also discarded, improving search relevance. The `vector_quality_threshold` config has been consolidated into `similarity_threshold` (default changed from `0.0` to `0.3`). (#1 | Medium | 4/12/2026 |
| v0.2.27 | ## What's Changed ### 🐛 Bug Fixes - Normalize RRF scores against theoretical max and use message order for fork divider (#1445) - Resolve duplicate key violation on concurrent file ingestion (#1444) - Find latest release with CLI binaries instead of blindly using latest (#1443) ## Upgrade Run `tale deploy` to update to this version. No manual migration required. ## Contributors @larryro --- **Full Changelog**: https://github.com/tale-project/tale/compare/v0.2.26...v0.2.27 | Medium | 4/12/2026 |
| v0.2.26 | ## What's Changed ### 🐛 Bug Fixes - Enforce team access control on prompt templates — users can now only view and modify prompt templates belonging to teams they are a member of (#1442) ## Upgrade Run `tale deploy` to update to this version. No manual migration required. ## Contributors @larryro --- **Full Changelog**: https://github.com/tale-project/tale/compare/v0.2.25...v0.2.26 | Medium | 4/12/2026 |
| v0.2.25 | ## What's Changed ### 🚀 Features - **Auto model selection**: The model selector now defaults to "Auto" mode, which automatically selects the best available model with fallback. Manual model overrides are preserved for 24 hours before reverting to Auto. (#1441) ### 🐛 Bug Fixes - **Expanded error categorization**: Added detection for authentication errors (401/403, invalid keys), model-not-found errors (404), and provider errors (5xx, overloaded). Known error types now show clean i18n messages | Medium | 4/12/2026 |
| v0.2.24 | ## What's Changed ### 💥 Breaking Changes - **OpenAI-compat API refactored from dual-mode to direct model gateway** (#1440) - **Agent mode removed**: The API no longer supports server-side agent execution (previously triggered when no `tools` param was sent). All requests now route directly to providers by model ID. - **`model` parameter changed**: Previously accepted agent slugs (e.g. `"chat-agent"`); now requires provider model IDs (e.g. `"anthropic/claude-sonnet-4.6"`). Use `GET /api/v1/ | Medium | 4/12/2026 |
| v0.2.23 | ## What's Changed ### Bug Fixes - **Fix fallback models not triggering on non-transient provider errors** — The model fallback chain now activates on provider-specific errors (401 auth failures, 404 model-not-found) in addition to transient errors (429, 5xx). A different fallback model on a different provider can now recover from configuration issues. Also fixes UI flicker (loading indicator disappearing, error messages flashing) between fallback retries by suppressing error cleanup during the | Medium | 4/12/2026 |
| v0.2.22 | ## What's Changed ### Features - Add citation support for context retrieval mode (#1437) — RAG and web search results now include source citations that are passed through to the agent response, enabling inline references in chat ### Bug Fixes - Fix nested button hydration error and dialog overflow (#1436) — resolve React hydration mismatch in team-multi-select and fix message info dialog overflow on smaller viewports ## Upgrade Run `tale deploy` to update to this version. No manual migration | Medium | 4/12/2026 |
| v0.2.21 | ## What's Changed ### 🐛 Bug Fixes - Fix citations display and source card file preview (#1435) - Add `aria-labelledby` to chat interface for WCAG 2.1 compliance (#1431) - Trim de-AT and de-CH locale files to only contain ß→ss overrides, removing duplicate entries ## Upgrade No manual migration required. Run `tale deploy` to update to this version. ## Contributors @larryro, @yannickmonney --- **Full Changelog**: https://github.com/tale-project/tale/compare/v0.2.20...v0.2.21 | Medium | 4/12/2026 |
| v0.2.20 | ## What's Changed ### 🚀 Features - Add MCP server management UI and backend (#1408) - Add audit log immutability, AI usage logging, and export (#1405) - Add per-model API key and base URL overrides (#1404) - Add model access RBAC governance policy (#1403) - Expose RAG citations in OpenAI-compatible API (#1402) - Surface OCR visibility with scanned page detection (#1401) ### 🐛 Bug Fixes - Polish teams UI and fix update payload (#1424) ### ⚡ Performance - Reduce TTFT with parallelization and | Medium | 4/11/2026 |
| v0.2.19 | ## What's Changed ### 🚀 Features - Add multi-team agent access control — agents can now be shared across multiple teams (#1399) - Add org-scoped chat sharing with direct link support (#1397) - Add team-based workspace chat filtering (#1394) - Configurable upload and retention policies (#1393) - Wire default model resolver and add governance tab (#1390) ### 🐛 Bug Fixes - Correct first-message layout shift on new thread creation (#1416) - Improve model validation and error surfacing for AI fai | Medium | 4/11/2026 |
| v0.2.18 | ## What's Changed ### 🚀 Features - Wire context length governance into context builder (#1388) - Wire prompt library dialogs into UI (#1386) - Mount canvas provider and render canvas pane (#1385) - Wire dictation button into chat input (#1384) - Wire message feedback into chat bubbles (#1383) - Wire feature flags resolver and add governance tab (#1375) ### 🐛 Bug Fixes - Improve error messages and add retry for chat failures (#1376) ### ♻️ Refactoring - Real-time form validation and legacy i | Medium | 4/11/2026 |
| v0.2.17 | ## What's Changed ### 🚀 Features - **File tools simplification** — auto RAG indexing on upload, removed parse/template operations for a cleaner file tool surface (#1407) - **PII policy governance** — new PII policy tab on the governance page with configurable detection patterns and mask/block modes (#1370) - **Configurable file retention policies** — granular retention settings for user and agent temporary files with per-category TTL (#1391) - **Model fallback mechanisms** — reusable ModelSele | Medium | 4/11/2026 |
| v0.2.16 | ## What's Changed ### 🚀 Features - Arena mode for side-by-side model comparison (#1348) ## Upgrade Run `tale deploy` to update to this version. No manual migration required. --- **Full Changelog**: https://github.com/tale-project/tale/compare/v0.2.15...v0.2.16 | Medium | 4/10/2026 |
| v0.2.15 | ## What's Changed ### 🚀 Features - Add BYOM (Bring Your Own Model) support and per-model configuration (#1319) ### 🐛 Bug Fixes - Prevent Swagger docs page from redirecting on click (#1329) - Refine toast icons, CSV header detection, and vendor bulk delete (#1321) ## Upgrade Run `tale deploy` to update to this version. No manual migration required. ## Contributors @Israeltheminer, @larryro, @yannickmonney --- **Full Changelog**: https://github.com/tale-project/tale/compare/v0.2.14...v0.2. | Medium | 4/10/2026 |
| v0.2.14 | ## What's Changed ### Features - Add chat export as PDF and Markdown (#1322) - Add budget limits per user/group (#1311) ### Bug Fixes - Add thread ownership checks to prevent unauthorized access (#1326) - Add submit button and auto-close to document upload (#1320) - Add success variant to toast notifications (#1317) - Add name field to vendor import form (#1315) - Improve member removal success toast message (#1316) - Add missing import guidance for customers and vendors (#1314) - Fix vendor l | Medium | 4/10/2026 |
| v0.2.13 | ## What's Changed ### Features - Add REST API with OpenAPI spec and API key authentication (#1295) - Add `tale auth reset-owner` CLI command (#1298) ### Bug Fixes - Update dependency @microlink/react-json-view to v1.31.18 (#1294) - Update dependency convex to v1.35.1 (#1278) ### Other Changes - Update oven/bun Docker tag to v1.3.12 (#1286) ## Upgrade Run `tale deploy` to update to this version. No manual migration required. ## Contributors @larryro --- **Full Changelog**: https://github.c | Medium | 4/10/2026 |
| v0.2.12 | ## What's Changed ### 🚀 Features - **CLI: Download progress indicator for upgrades** — The `tale upgrade` command now shows real-time download progress (bytes downloaded, percentage) in TTY terminals. Also removes the compiled binary name guard (allowing dev builds to self-upgrade) and detects dev builds for automatic upgrade. (#1292) ### 📝 Other Changes - **deps:** Update Bun to v1.3.12 (#1285) ## Upgrade Run `tale deploy` to update to this version. No manual migration required. ## Cont | Medium | 4/10/2026 |
| v0.2.11 | ## What's Changed ### 🚀 Features - Add chat sharing and forking — share conversations via link and fork shared chats to continue them (#1282) - Add inline source citations for RAG and web search results (#1279) - Filter model selector to only show chat-tagged models (#1277) - Add intelligence icons and tooltips to model selector (#1259) ### 🐛 Bug Fixes - Expose raw error messages and flatten tool schemas for OpenAI provider compatibility (#1276) ### 📦 Dependencies - Update mysql2 to v3.21. | Medium | 4/10/2026 |
| v0.2.10 | ## What's Changed ### 🚀 Features - Add chat message editing and conversation branching (#1271) - Implement WCAG 2.1 AA accessibility standards (#1266) - Add LLM response caching for agents (#1260) - Enforce mandatory governance system prompt (#1257) - Implement trusted headers authentication middleware (#1252) - Add OpenAI Chat Completions API compatibility layer (#1251) ### 🐛 Bug Fixes - Fix agent save for delegation and visibility settings (#1245) - Update dependency @sentry/tanstackstart- | Medium | 4/9/2026 |
| v0.2.9 | ## What's Changed ### 🚀 Features - Chat archive, file upload improvements & provider edit dialog (#1236) ### 🐛 Bug Fixes - Add @modelcontextprotocol/sdk as dependency and external package (#1247) - Sanitize automation creation error messages (#1240) - Show friendly error for duplicate folder names (#1243) - Fix agent deletion silently swallowing errors (#1244) - Extract semver robustly from release tags in upgrade check (#1241) ### 📦 Dependencies - Update mysql2 to v3.21.0 (#1246) - Update | Medium | 4/9/2026 |
| v0.2.8 | ## What's Changed ### 🚀 Features - **Governance framework** — policies, budget enforcement, cost estimation, PII detection/masking, retention cleanup, upload enforcement, and feature enforcement (#1195 and related commits) - **MCP server integration** — connect external MCP servers as agent tools with OAuth2 support (#1192) - **Arena mode for chat** — side-by-side model comparison with verdict bar (#1195) - **Canvas pane** — render code, HTML, Markdown, and Mermaid diagrams inline in chat (#11 | Medium | 4/9/2026 |
| v0.2.7 | ## What's Changed ### 🚀 Features - Add AI image generation integration with approval flow (#1114) - Add `conversation_read` agent tool (#1107) - Add model selection to agent creation dialog (#1109) - Add pagination to product and customer tables (#1108) - Add bulk select and delete for products and customers (#1098) - Add manual product entry UI (#1092) - Add status field to product and customer edit forms (#1093) - Add logout confirmation dialog (#1095) - Add dedicated email column to custome | Medium | 4/8/2026 |
| v0.2.6 | ## What's Changed ### 🚀 Features - **Crawler**: Add BFS fallback for URL discovery & distinguish org not-found on dashboard (#1014) ### 🐛 Bug Fixes - **Crawler**: Replace crawl4ai URL discovery and improve website registration (#1013) - **Crawler**: Add PATCH endpoint for scan interval updates (#1010) - **Platform**: Preserve user message content integrity (#1009) - **Platform**: Prevent duplicate identifiers and protect system agents (#1003) - **UI**: Restore designs lost during merge confl | Medium | 4/7/2026 |
| v0.2.5 | ## What's Changed ### 🚀 Features - Optimize Docker containers — 60% total image reduction (~10 GB → ~4 GB) + container e2e tests (#958) - Archive chat, sticky notes, API keys & team management (#970) ### 🐛 Bug Fixes - Harden sign-up flow, session handling, and crawler DB init (#996) - Remove redundant HSTS header and default TLS email in proxy (#995) ### 📝 Other Changes - Update github/codeql-action to v4 (#994) ## Upgrade No manual migration required. Run `tale deploy` to update. ## Co | Medium | 4/7/2026 |
| v0.2.4 | ## What's Changed ### 🚀 Features - Stream container logs during deploy for real-time visibility into startup progress (#989) - Per-section auto-save for providers and unsaved changes guard (#992) ### 🐛 Bug Fixes - Fix race condition that prevents database creation on fresh deploy (#993) - Fix ownership of synced files after `docker cp` to prevent EACCES errors when the app user writes to synced directories (#990) - Drop privileges at entrypoint to fix data directory permissions on Docker vol | Medium | 4/6/2026 |
