v0.16.7

Letta Server 0.16.7 Release Notes

173 commits since 0.16.6 | Released March 31, 2026

Highlights

Self-hosted users: this is a big upgrade. The default global context window is raised from 32k to 128k, the context window reset bug (LET-7991) is fixed, and compaction has been overhauled. If you've been running curl commands to patch your config after every ADE load, most of that pain should be gone.

Breaking Changes

• Block limits are no longer enforced -- block limit validation has been deprecated and removed from the git memory sync path (#9977, #9983). Blocks can now grow freely. If you were relying on limits to cap per-turn cost, you'll need to manage block size via other means.

Context Window & Compaction (21 fixes)

The biggest category of fixes. Self-hosted users were hit hardest by these.

• Global context window default raised from 32k to 128k (#9993) -- self-hosted servers no longer default to 32k for unknown models
• Context window preserved on conversation model override (LET-7991, #9986) -- the bug where non-default conversations fell back to 32k is fixed
• Compaction overflow fixes (#9897) -- addresses the double-compaction and runaway compaction loops
• Compaction model resets on agent model change (#10031) -- switching your agent's model no longer leaves the old summarizer model behind
• Summarizer prompt improved (#10314) -- now remembers plan files, GitHub PRs, and other structured content during summarization
• BYOK summarization fixed (#10152) -- summarizer provider fallback no longer fires for BYOK requests
• Better error surfacing -- context window exceeded errors now have descriptive messages (#10135, #10171), and system prompt size warnings during compaction (#10058)

Gemini (2 fixes)

• thought_signature preserved on function calls without reasoning (LET-8166, #10237) -- the bug blocking all Gemini 2.5+/3.x multi-turn tool calling is fixed
• Streaming interface crash fixed (#10306) -- self.model now initialized in SimpleGeminiStreamingInterface constructor (LET-8129)

Memory & memfs (10 fixes, 4 features)

• available_skills block no longer duplicates in system prompt (#10006, #10011, #10021) -- three separate fixes for the skills block multiplying and inflating context (LET-8013)
• Git memory sync deferred until stream close (#9951) -- reduces mid-stream sync failures
• System prompt recompiles on agent creation with git memory (#9950) -- new git-enabled agents no longer start with empty compiled context
• Projection-style git memory rendering (#10211) -- new rendering approach for memfs content in system prompts
• Manual block edits via API trigger recompile (#9775) -- no more stale context after API block updates
• Conversation recompile endpoint (#9848) -- POST /v1/conversations/{id}/recompile is now available

Conversations (7 features)

• Conversation forking (#10234, #10263) -- fork conversations with shared message history, including the default conversation
• Sort conversations by last_message_at (#10190)
• Idempotent conversation streaming (#10147) -- OTID-based retry safety
• Request-scoped system overrides (#10227) -- per-request system prompt modifications

Streaming & Reliability (13 fixes)

• OTID retry hardening (#10229, #10209) -- stream resume with backoff, race condition fixes
• Conversation lock released earlier (#10203) -- reduces contention on concurrent requests
• Better error messages (#10207) -- known LLM errors now surface descriptive messages instead of generic failures
• BYOK error tagging (#10204, #10311) -- errors now include is_byok flag for debugging
• stream_incomplete diagnosis (#10033) -- BaseException catching to identify root causes

Model Support (14 features, 20 fixes)

• GPT-5.4 -- full support including mini, nano, and fast variants (#9798, #10043)
• GLM-5 -- GLM-5, GLM-5.1, GLM-5 Turbo, GLM-4.7 (#10317, #10285, #9994)
• MiniMax M2.7 (#10093)
• Baseten -- added as provider with full frontend integration, serverless auto mode, reasoning support (#10250, #9846, #9998)
• Fireworks (#9780) and zAI coding provider (#10064)
• Opus 4.6 / Sonnet 4.6 -- adaptive thinking tokens no longer incorrectly capped (#9795)
• OpenAI proxy cleanup -- extra fields removed (#9949), parallel tool calling supported (#9879)

Security

• Local filesystem access blocked via ImageContent bypass (#3256, #10329) -- file:/// URLs in images are now rejected
• Internal MCP server targets blocked (#10009)
• SECURITY.md added (#3228)

Infrastructure

• Readiness enforcement scaffold (M1-M3 metrics pipeline) -- request pressure, DB pool, SSE lifecycle, event loop lag monitoring
• Multi-agent tools moved to less privileged execution environment (#9779)
• Subagent agents auto-hidden on create (#10096)
• WebSocket transport for OpenAI Responses API (#9841)

---

For self-hosted users upgrading from 0.16.6: This release addresses the majority of issues reported in the community over the past month. The context window default change alone (#9993) eliminates the most common source of "everything breaks when I open ADE" complaints.

Full Changelog: 0.16.6...0.16.7

v0.16.6

Highlights

• Expanded Conversations API support for default conversation / agent-direct mode.
• New conversations now initialize with a compiled system message at creation time.
• Fixed model_settings.max_output_tokens default behavior so it does not silently override existing max_tokens unless explicitly set.

Conversations API updates

• Added support for conversation_id="default" + agent_id across conversation endpoints (send/list/cancel/compact/stream retrieve).
• Kept backwards compatibility for conversation_id=agent-* (deprecated path).
• Added lock-key handling in agent-direct flows to avoid concurrent execution conflicts.

Conversation/system-message behavior

• Conversation creation now compiles and persists a system message immediately.
• This captures current memory state at conversation start and removes first-message timing edge cases.

Model/config updates

• Added model support for:
  • gpt-5.3-codex
  • gpt-5.3-chat-latest
• Updated defaults:
  • context window default: 32k → 128k
  • CORE_MEMORY_BLOCK_CHAR_LIMIT: 20k → 100k
• Anthropic model settings now allow effort="max" where supported.
• Gemini request timeout default increased to 600s.

Memory / memfs updates

• Git-backed memory frontmatter no longer emits limit (legacy limit keys are removed on merge).
• Skills sync now maps only skills/{name}/SKILL.md to skills/{name} block labels.
• Other markdown under skills/ is intentionally ignored for block sync.
• Memory filesystem rendering now includes descriptions for non-system/ files and condenses skill display.

Reliability and compatibility fixes

• Added explicit LLMEmptyResponseError handling for empty Anthropic streaming responses.
• Improved Fireworks compatibility by stripping unsupported reasoning fields.
• Improved Z.ai compatibility by mapping max_completion_tokens to max_tokens.

Full Changelog: 0.16.5...0.16.6

v0.16.5

What's Changed

• chore: bump version 0.16.5 by @carenthomas in #3202

Full Changelog: 0.16.4...0.16.5

v0.16.4

What's Changed

• fix: update gh templates by @cpacker in #3155
• chore: release 0.16.3 by @sarahwooders in #3158
• chore: bump v0.16.4 by @carenthomas in #3168

Full Changelog: 0.16.2...0.16.4

v0.16.2

What's Changed

• docs: update README.md by @cpacker in #3110
• Update contributing.md with corrected local setup steps by @neversettle17-101 in #3123
• chore: bump version 0.16.2 by @carenthomas in #3140

New Contributors

• @neversettle17-101 made their first contribution in #3123

Full Changelog: 0.16.1...0.16.2

v0.16.1

What's Changed

• Correct provider name for openai-proxy in LLMConfig by @SootyOwl in #3097
• chore: bump v0.16.1 by @carenthomas in #3107

Full Changelog: 0.16.0...0.16.1

v0.16.0

What's Changed

• Updated readme with actual argument by @Godofnothing in #3083
• fix: Implement architecture-specific OTEL installation logic by @SootyOwl in #3061
• chore: bump v0.16.0 by @carenthomas in #3095

New Contributors

• @Godofnothing made their first contribution in #3083
• @SootyOwl made their first contribution in #3061

Full Changelog: 0.15.1...0.16.0

v0.15.1

What's Changed

• chore: bump version 0.15.1 by @carenthomas in #3082

Full Changelog: 0.15.0...0.15.1

v0.15.0

What's Changed

• Add context windows for grok-4 models by @runtimeBob in #3043
• chore: bump version 0.15.0 by @carenthomas in #3077

New Contributors

• @runtimeBob made their first contribution in #3043

Full Changelog: 0.14.0...0.15.0

v0.14.0

What's Changed

• chore: bump version 0.14.0 by @carenthomas in #3067

Full Changelog: 0.13.0...0.14.0