Changelog

​

Added

• Session allowlist on the Dialectic and representation via a constrained filters body on POST /peers/{peer_id}/chat and /representation, supporting only the session_id key (a session id, a bare list, or {"in": [...]}). Unsupported keys and shapes are rejected with 422 rather than silently ignored, it composes with session_id (which must be included in the allowlist when both are given), and it is capped at 1,000 sessions per request. Enforcement is uniform and fail-closed at every recall chokepoint: scoped conclusion recall is restricted to level == "explicit" (dream-derived conclusions carry a single session_name but are synthesized across all sessions, so that stamp can’t be scoped on), get_reasoning_chain is unavailable under an allowlist, and an empty allowlist short-circuits to empty results everywhere. Workspace keys pass the allowlist as-given; peer-scoped JWTs must be an active member of every allowlisted session (403 otherwise) (#882)
• Bare-list membership sugar in the filter DSL: {"session_id": ["s1", "s2"]} is now shorthand for {"session_id": {"in": [...]}} on regular columns generically. JSONB metadata columns are excluded and keep containment semantics. Strictly additive, since a bare list on a regular column previously compiled to a type-mismatched equality that matched nothing (#881)
• Optional structured outputs on the Dialectic: response_format (a JSON Schema with root type object) on peer chat makes content a JSON string conforming to that schema. Only a conservative subset of JSON Schema is supported, with DoS guards and non-recursive $ref support (#896)
• Combined tool calling and structured output in the LLM transport layer, with per-backend request shaping: OpenAI routes tool-carrying structured requests through create() with an explicit json_schema response format (parse() 500s on non-strict function tools), Anthropic skips the { JSON prefill when tools are present so tool_use blocks stay reachable, and Gemini injects a schema instruction into the final turn instead of using native response_schema (rejected alongside function calling before Gemini 3). All backends skip structured-output parsing on tool-call turns, which carry no consumable content (#907)
• card_refresh dream type: a lightweight dream that runs only the peer-card update, for event-driven refreshes such as membership changes and cold starts. Handled by a new CardRefreshSpecialist restricted to get_recent_observations, search_memory, and update_peer_card (no observation-mutating tools) with a tool-iteration cap of min(6, DREAM.MAX_TOOL_ITERATIONS). POST /v3/workspaces/{workspace_id}/schedule_dream accepts dream_type=card_refresh plus a rebuild flag, which omits the existing card from the prompt so the specialist rebuilds it solely from observations present in the collection. Card refreshes never advance the omni dream guard pair (last_dream_at / last_dream_document_count) (#883)
• Full-fidelity LLM trace stream, with Langfuse as one projection over it: each call is captured once (CapturedLLMCall) and fanned out to a CloudEvents trace stream (llm.call.traced / trace.content) and a Langfuse exporter, both reconstructing trace → run → step → generation from the same source of truth. Adds TELEMETRY_TRACE_PAYLOADS_ENABLED (default false), TELEMETRY_TRACE_MAX_BYTES (default 262144, per-message cap with oversized content clipped), TELEMETRY_TRACE_PURPOSES (JSON list of CallPurpose values; empty means all), and LANGFUSE_EXPORTER_MODE (exporter by default; inline is kept for one release for side-by-side validation). Embedding calls are traced, dreamer branches nest under one dream trace, tool calls become spans under their step, and high-volume events are sampled deterministically. TRACE_ENDPOINT is dropped (#845)
• Redis Cluster support via CACHE_CLUSTER (for example GCP Memorystore for Redis Cluster), alongside a new CACHE_LOCK_WAIT_CHECK_INTERVAL_SECONDS (#905)
• EMBEDDING_MODEL_CONFIG__MAX_BATCH_SIZE caps texts per embedding request for OpenAI-compatible providers with smaller limits than OpenAI’s, such as DashScope text-embedding-v4 (10) and Alibaba Bailian qwen3.7-text-embedding (20). When unset, native provider defaults are preserved (OpenAI 2048, Gemini 100) (#983)
• Per-request provider timeouts via provider_params.timeout on any model config, validated at config load so a bad value fails at startup with the exact config path instead of surfacing per-request as a retried 500. Good values normalize to float seconds; Gemini’s is converted to milliseconds (#832)
• RepresentationCompletedEvent now reports deduplication counts: exact_dup_in_batch_count, exact_dup_existing_count, semantic_dup_rejected_count, and semantic_dup_replaced_count (#910)
• OAuth discovery for MCP clients: the MCP worker serves /.well-known/oauth-protected-resource (RFC 9728) without auth so clients can discover the authorization server, and a 401 now carries WWW-Authenticate: Bearer resource_metadata="..." (exposed cross-origin) to start the flow (#923)
• Prometheus metrics for the immediate-embed fast path: tasks shed because EMBEDDING_MAX_PENDING_EMBED_TASKS was reached, and the current in-flight task count (#892)
• Docs: a detailed system architecture diagram, a Codex integration guide (#879), a structured-outputs page (#896), a section on filtering conclusions by reasoning level (#851), a health-check endpoint reference, and SDK updates (#867)

​

Changed

• Breaking config change: DERIVER_REPRESENTATION_BATCH_MAX_TOKENS is split into two settings that were previously conflated — DERIVER_REPRESENTATION_BATCH_WORK_UNIT_TARGET_TOKENS (default 512), the producer-side minimum a work unit accumulates before the deriver claims it, where 0 disables the gate; and DERIVER_REPRESENTATION_BATCH_TARGET_INPUT_TOKENS (default 1024), the consumer-side maximum context-window tokens per deriver LLM call. Deployments setting the old name must migrate (#889)
• The immediate-embed fast path now applies backpressure: EMBEDDING_MAX_PENDING_EMBED_TASKS (default 50) caps in-flight embed tasks, and once saturated, message creation skips the fast path entirely and the reconciler embeds on its next cycle. 0 disables the fast path (#892)
• Explicit-level documents are now kept session-pure, so memory can be built by copying explicit documents between collections. Enforcement refuses rather than rewrites: create_documents rejects explicit documents with a null session_name, exact dedup keys on (content, level, session-for-explicit), semantic dedup scopes candidate search to the same level and — for explicit documents — the same session, and the generic create_observations tool rejects level='explicit' outside message-ingestion (deriver) context. Derived levels keep cross-session consolidation (#883)
• Sentry’s before_send filter is centralized as default_before_send in src/telemetry/sentry.py instead of living only in the API’s main.py, so the deriver gets the same non-actionable-exception filtering. All Sentry events also carry a namespace tag for correlation (#934, #870)
• The minimal deriver’s extraction examples no longer teach inferences its own output schema forbids. The EXAMPLES block demonstrated deriving a specific birthday from “I just had my 25th birthday last Saturday”, deriving residence from a single visit (“I took my dog for a walk in NYC” → “alice lives in NYC”), and a ”+ general knowledge” deductive output the deriver has no channel for. The replacements stay inside the schema’s contract and teach the boundary: the dog/NYC message is kept and shown extracting correctly, and a separate example shows “lives in NYC” is valid when actually stated (#985)
• Dreamer specialists are instructed not to output summaries (#894)
• session_name is deprecated for scoping in favor of the session allowlist. It is not removed and not aliased: it also pins the query to one session, bypasses observer scoping, and drives session-history injection into the dialectic prompt, so it has no drop-in replacement (#882)
• The MCP worker no longer requires the X-Honcho-User-Name or X-Honcho-Assistant-Name headers (#923)

​

Fixed

• Session scoping was applied to only one of the three working-representation query paths: session_name reached the recent-documents query, but the semantic and most-derived paths ignored it, so limit_to_session leaked cross-session conclusions into perspectives. The allowlist is now threaded uniformly through all three paths and pushed down to pgvector and external vector stores (#881)
• Empty membership lists failed open in the vector-store filter builders, silently widening scope: LanceDB dropped empty IN clauses and Turbopuffer emitted a bare In [] with undocumented semantics. Both now emit an explicit always-false predicate, and _build_filter_conditions checks is not None rather than truthiness so an empty list is no longer treated like None (#881, #882)
• Session-scoped CRUD helpers ignored the session allowlist entirely, so a caller could read a session the allowlist forbids. The API routes guarded this with a 422, but the dialectic tools call these CRUD functions directly and bypassed it. _semantic_search_messages (covering search_messages and search_messages_temporal), grep_messages, get_messages_by_date_range, get_recent_history, and get_observation_context now return [] when session_name is set and outside the allowlist (#882)
• The cache client logged the full Redis URL — including the password — at INFO and WARNING on every connection attempt and failure, exposing the live credential in container logs and downstream aggregation. Credentials are now redacted across userinfo, the ?password= (redis-py) and ?secret= (cashews) query params, scheme-less URLs whose password is invisible to .port/.password parsing, and malformed URLs, whose fallback previously echoed the raw input verbatim (#869)
• A top_k of 0 reached the vector store, where Turbopuffer rejects it with a 400 (top_k must be between 1 and 10000). A non-positive top_k now returns [] before the embedding call, and the semantic budget floors at 1 so an explicitly requested search isn’t silently allocated zero (#970)
• Gemini clients had no HTTP timeout, so a stalled socket wedged the deriver worker’s uvloop event loop, which the in-process reconciler shares. A 10-minute timeout is now set on both the Gemini LLM client and the Gemini embedding client (#903)
• Dreamer conclusions were dated to ingestion time rather than their latest source observation, and their timestamps are now normalized (#890)
• Langfuse I/O annotation was gated on LANGFUSE_PUBLIC_KEY instead of langfuse_inline_enabled, so in the default exporter mode it called update_current_generation() with no active span — logging “No active span in current context” roughly 14 times per dialectic run and building throwaway model_dump payloads on every LLM call. Separately, AgentToolSummaryCreatedEvent hardcoded run_id="deriver" / iteration=0, polluting run_id grouping in the CloudEvents stream with a phantom run; both fields are now optional and the resource id is keyed on message_id:summary_type (schema_version 2 → 3) (#845)
• Assistant tool calls were dropped from the captured trace stream for OpenAI and Gemini: build_captured_messages read only {role, content, tool_call_id}, but those providers keep tool calls outside content, so replayed tool-call turns landed as empty content and Gemini lost its text and tool results entirely. Tool calls are now normalized per provider into a unified tool_calls field and folded into the content hash. Gemini’s thought_signature is bytes, so model_dump(mode="json") raised UnicodeDecodeError inside emit_trace, silently dropping whole tool-calling iterations from the trace stream (billing and Langfuse were unaffected); it is now base64-encoded on the telemetry path while replay keeps the raw bytes (#845)
• EmbeddingClient.encoding forced full client construction, raising “OpenAI API key is required” even though tiktoken needs no credentials. The document dedup tie-break only needs .encoding for token counting, so any test hitting that path failed in environments without embedding keys — notably CI for pull requests from forks. The encoding is now resolved from the configured model directly, falling back to cl100k_base, and the underlying client’s encoding is reused only when it has already been constructed (#955)
• The Docker build failed under Podman because the uv build inputs weren’t copied (#878)
• LanceDB was installed on macOS Intel, where it doesn’t work. A PEP 508 marker excludes darwin/x86_64 and the LanceDB vector-store import is wrapped so a misconfiguration surfaces as a clear config error (#496)
• Prompt checks requiring the literal token “json” for json_object mode are now satisfied in lowercase (#887)
• Reverted an unintended RepresentationCompletedEvent schema-version increment
• Documented preinstalling pgvector as a privileged role for deployments where the DB_CONNECTION_URI role deliberately cannot create extensions (managed Postgres, Kubernetes operators, NixOS). CREATE EXTENSION IF NOT EXISTS vector does not help there, because Postgres checks the privilege before checking whether the extension exists. Docker Compose is unaffected, since the bundled stack connects as the postgres superuser (#984)

​

Added

• api_request_duration_seconds Prometheus histogram tracking per-route request latency, labeled by method and endpoint (#837)
• LLM provider_params passthroughs (extra_body / extra_headers / extra_query) are now forwarded to the underlying provider transport across all backends, with shape validation that rejects non-mapping values (#821)
• structured_output_mode model-config option to use json_object mode for OpenAI-compatible providers that lack native Structured Outputs support (used by the deriver) (#820)
• OpenRouter app-attribution headers (HTTP-Referer / X-Openrouter-Title) are now sent on OpenAI-compatible clients when the configured base URL is OpenRouter, so requests are attributed to “Honcho” in OpenRouter’s dashboard (#805)
• Langfuse traces are now tagged with user and session IDs for easier trace filtering (#814)
• DERIVER_REPRESENTATION_BATCH_MAX_AGE_SECONDS (default 1800s) lets sub-threshold representation work units flush once their oldest unprocessed queue item ages out. Set it to 0 to keep the legacy behavior where sub-threshold tails wait indefinitely unless DERIVER_FLUSH_ENABLED=true (#826)
• Conclusion responses now include a level field (explicit, deductive, inductive, contradiction); list/query endpoints support filtering by level via filters, with reserved filter keys protected from being overridden by user-supplied filters (#851)

​

Changed

• Peer-scoped JWTs now get read-only access to the sessions their peer is an active member of (session context, summaries, peers, their own per-session config, search, and message reads). Session-scoped JWTs remain confined to their session and cannot reach peer routes (#679)
• Compacted Honcho’s log output, with guarded ms/s metric formatting that falls back to a plain string for non-numeric values (#836)
• Sentry now drops noisy infra/scrape transactions: the reconciler opens a transaction only once a batch has rows (idle cycles emit none), and a traces_sampler returns 0.0 for /metrics, /health, /openapi.json, /docs, /redoc, and the deriver metrics server. SENTRY.TRACES_SAMPLE_RATE still governs real traffic (#834)

​

Fixed

• Peer- and session-scoped JWTs were effectively workspace-scoped: authorization walked the route’s declared scope and fell through to a workspace match, so a {w, p: alice} token could act on any peer in the workspace. JWTs are now authorized by their narrowest claim and never widen to workspace access (#679)
• The keys API now rejects creating a peer- or session-scoped key without a workspace. Such keys were minted successfully but failed verification on every request (#679)
• Agent-supplied observation IDs carrying the display-format id: prefix are now normalized (prefix and trailing whitespace stripped) before source_ids are stored and on get_reasoning_chain lookups, fixing corrupted provenance links and broken reasoning-chain traversal (#795)
• Fixed a create_tree keyword-argument mismatch in the Dreamer’s surprisal tree construction (#749)
• Providers that omit output-token counts (observed with Gemini on tool-loop completions) returned output_tokens=None, which raised a Pydantic validation error that aborted the call and crashed the Dreamer’s induction phase before inductive conclusions were persisted. None is now coerced to 0 so token accounting degrades gracefully (#809)
• Document creation now performs exact (case-insensitive, whitespace-trimmed) content deduplication before the existing semantic dedup step: exact duplicates within a batch collapse to a single insert, and an exact match against a live document reinforces it (atomic times_derived increment) instead of creating a new row (#861)
• The OpenAI backend passed tool_choice through raw while the Anthropic and Gemini backends translate Honcho’s canonical vocabulary to their native form, so on a mixed-provider fallback chain (for example Gemini primary → OpenAI backup) a canonical "any" reached OpenAI unchanged and was rejected as an invalid param. The OpenAI backend now converts it, mirroring the others: any/required → required, auto/none pass through, and a tool-name string or {"name": ...} dict becomes a function selection (#850)
• Langfuse @observe auto-capture serialized every argument of honcho_llm_call_inner into the generation span input, including client_override (a live AsyncOpenAI/genai client) and selected_config (which carries api_key). Auto-capture deep-copied the client into a half-constructed object whose teardown raised (AsyncHttpxClientWrapper ... no attribute '_state' on OpenAI, flooding stderr; BaseApiClient ... no attribute '_http_options' on Gemini), and it leaked ModelConfig.api_key into traces. Capture is now an explicit allowlist: capture_input/capture_output are disabled and curated, serializable input and output are stamped instead, with tuning knobs surfaced as model_parameters via a secret-bearing denylist and per-call token usage mirrored as usage_details (#849)

​

Added

• Messages are now embedded via a background task rather than blocking API request
• Read-only DB session mode (get_read_db / tracked_db(..., read_only=True)) so reads don’t hold a transaction open across the work
• CORS_ORIGINS env var to configure CORS allowed origins without editing source; defaults match the prior hardcoded list, so self-hosted deployments behind custom domains can whitelist their frontend (#697)
• scripts/generate_jwt.py — utility for minting scoped or admin Honcho JWTs (--admin, --workspace/--peer/--session, --expires with human-friendly durations, --print-only) without calling the keys API (#757)
• STALE_WORK_UNIT_CLEANUP_INTERVAL_SECONDS (default 60s) — minimum jittered spacing between deriver stale-work-unit cleanup runs, so cleanup no longer runs on every seconds-scale poll (0.0 keeps the legacy every-poll behavior) (#773)

​

Changed

• Optimized the deriver and dreamer prompt cache prefixes to improve prompt-cache hit rates (#806)

​

Fixed

• times_derived is now properly reinforced when a duplicate conclusion is detected. It had been pinned at 1 for nearly every conclusion (the reject-new branch dropped the increment and the new-wins branch reset the count to 1), so ORDER BY times_derived DESC fell back to arbitrary heap order and froze stale conclusions to the front of injected context. Reinforcement is now an atomic increment and both most-derived queries gained a created_at DESC recency tiebreaker (#768)
• Webhook creation now correctly rejects private/internal IP addresses (#793)

​

Changed

• Connection acquisition is now a single attempt with no server-side retry, on a vanilla AsyncSession. A new DB_CONNECT_TIMEOUT_SECONDS (default 2s) bounds the attempt so a saturated or unreachable pooler fails fast instead of holding a client connection open to re-knock. A saturated DB now surfaces to the caller — the API returns an error and the deriver backs off and retries on a later poll — which lets the pooler drain rather than amplifying saturation.

​

Added

• Deriver poll jitter so instances that start together don’t poll in lockstep: DERIVER_POLLING_STARTUP_JITTER_SECONDS (random delay before the first poll, default 30s) and DERIVER_POLLING_JITTER_RATIO (±fraction applied to every poll sleep, default 0.5). Both disable at 0.0; the underlying backoff schedule is unchanged.

​

Removed

• Reverted the connection-checkout retry and HonchoAsyncSession custom session introduced in 3.0.8. Removed the DB_CONNECTION_RETRY_ENABLED / DB_CONNECTION_RETRY_MAX_DELAY_SECONDS / DB_CONNECTION_RETRY_BACKOFF_INITIAL_SECONDS / DB_CONNECTION_RETRY_BACKOFF_MAX_SECONDS settings, the db_connection_acquisitions{outcome=...} Prometheus counter, and the db.pool.acquire Sentry span. Alerting built on db_connection_acquisitions should migrate to db_pool_connections / db_queries_in_flight.

​

Added

• Connection-checkout retry with bounded exponential backoff (tenacity) on get_db/tracked_db: transient transaction-pooler (Supavisor) rejections — SQLAlchemy TimeoutError and OperationalError — now retry with backoff instead of surfacing as 500s under client-connection saturation. Gated by DB_CONNECTION_RETRY_ENABLED with configurable delay/backoff knobs; ~10s default budget (#758)
• HonchoAsyncSession — a lazy AsyncSession that checks out its pooled connection (with retry) on the first DB-touching call rather than at construction. Request handlers doing non-DB work (embedding, file, LLM) before their first query no longer pin a pooler connection across it. Only the checkout is retried; the statement still runs exactly once, so writes are never duplicated (#758)
• Adaptive deriver queue polling: the poll interval backs off when the queue is idle or erroring (base → max, doubling each cycle) and snaps back to base the moment work is claimed, cutting steady-state query load against the DB. Gated by DERIVER_POLLING_BACKOFF_ENABLED with configurable max/multiplier (#758)
• New Prometheus db_pool_connections gauge (checked_out / checked_in / size / overflow), labeled api|deriver, registered in both the API lifespan and the deriver metrics server (#758)
• New Prometheus db_connection_acquisitions{outcome=ok|retried|exhausted} counter — the alertable early-warning signal that connection checkouts are retrying through pooler rejection, before requests start failing (#758)
• New Prometheus db_queries_in_flight gauge — statements actually executing on the wire (via SQLAlchemy cursor-execute events). Paired with checked_out, the gap reveals connections held but parked (the “idle in transaction during an external call” antipattern). Gated on METRICS.ENABLED for zero overhead when off (#758)
• Explicit SqlalchemyIntegration in both the API and deriver Sentry inits; connection acquisition wrapped in a db.pool.acquire span with live pool stats captured on retry exhaustion (#758)

​

Changed

• Default POOL_TIMEOUT lowered to 5s, with validation that it stays under the connection-retry budget when a pooled (non-null) POOL_CLASS is configured; config.toml.example and the v2/v3 configuration docs updated to match (#758)
• HonchoAsyncSession wraps every DB-touching session method (execute / scalar / scalars / flush / merge / refresh / commit / get / get_one / stream / stream_scalars / delete) so the lazy-checkout-with-retry guarantee has no holes; the acquired flag resets on close()/reset() so a reused session re-acquires on next use (#758)

​

Fixed

• Roll the session back on a retryable checkout failure before retrying — a failed autobegin could otherwise leave it pending-rollback, making the next connection attempt raise instead of cleanly re-checking-out (#758)
• Guard DBPoolCollector.collect() so a pool-read/import hiccup can’t raise and abort the entire /metrics scrape (Prometheus drops all metrics if any collector raises) (#758)
• Clamp the pool overflow gauge to ≥ 0 (it could report negative before the pool fills) (#758)
• Removed a double-sleep in the deriver idle poll so the backoff cap is a true cap rather than 2× (#758)

​

Added

• New src/llm/ package as the single owner of provider runtime: clients, backends, history adapters, tool loop, request builder, credentials, and caching policy (#459)
• New cloudevent LLMCallCompletedEvent (llm.call.completed) fires once per provider hit with full cost-attribution context: transport/provider_label, model, token counts with cache breakdown, finish_reason, outcome, retry/fallback state, duration, tool-call shape, streaming flag, and agent correlation (run_id + iteration) (#637)
• RepresentationCompletedEvent now carries total_input_tokens for full-trace cost attribution; per-emitter honcho_version injection; deterministic per-run_id high-volume sampler via TelemetrySettings.HIGH_VOLUME_SAMPLE_RATE (#637)
• Deriver custom instructions: per-workspace/peer guidance threaded into the deriver prompt with a MAX_CUSTOM_INSTRUCTIONS_TOKENS budget (default 2000); deriver MAX_INPUT_TOKENS raised 23000 → 25000 (#609)
• Configurable embedding dimensions: EMBEDDING_MODEL_CONFIG__DIMENSIONS_MODE (auto/always/never) controls whether OpenAI dimensions= is forwarded (#678)
• New honcho-cli package — Python CLI for inspecting and managing peers, sessions, and configuration against a Honcho deployment (#424)
• HONCHO_API_URL env var support in the MCP Worker for self-hosted deployments (#575)
• API ID max_length increased from 100 to 512 across WorkspaceCreate, PeerCreate, and SessionCreate to align with the DB schema (#684)
• AttemptPlan dataclass pins per-retry provider selection across stream-final retries so streaming doesn’t bounce back to primary after the tool loop has settled on fallback (#459)
• Gemini JSON-schema sanitizer for function_declarations — strips keywords Gemini’s validator rejects while preserving semantics for other backends (#459)

​

Changed

• All LLM orchestration moved out of src/utils/clients.py into src/llm/ with modules split by responsibility (#459)
• Default ModelConfig factories (deriver, summary, dreamer specialists, dialectic levels) normalized with no extra parameters set by default; operators add transport/thinking overrides explicitly (#459)
• OpenAI reasoning-model routing widened to cover gpt-5.x and o1/o3/o4 — these models receive max_completion_tokens instead of max_tokens (#459)
• Peer card prompts reframed as stable identity markers; induction specialist now opts out of peer card writes so only deduction touches the card (#686)
• Vector store queries no longer fetch embedding vectors — only document metadata is returned, reducing payload size and DB load (pgvector, lancedb, turbopuffer) (#682)
• Langfuse trace metadata now includes namespace, model, and provider so traces can be filtered by deployment slice (#565)
• Deriver: model-aware tokenizer (replaces the previously hardcoded encoding) and explicit guard on empty message content (#647)
• Dialectic level defaults now merge correctly with per-level overrides (#656)
• Default dialectic tool choice switched to auto (#630)
• Vector sync given a substantial retry budget to tolerate transient embedding provider outages (#604)
• AgentToolConclusionsDeletedEvent payload now carries levels (#612)
• Turbopuffer: InternalServerError caught and surfaced as a warning rather than a hard failure; vector store sync errors downgraded to warnings (#561)

​

Fixed

• reverse query parameter is now honored on the v3 workspace list, peer list, workspace-scoped session list, and peer-scoped session list. Honcho SDKs at 2.1.0+ were already sending reverse=true for these routes but the server silently ignored it. Ties on created_at now fall back to the internal nanoid id for stable ordering across pages (#685)
• LLM client factories now receive base_url from LLMSettings for default providers — operators pointing at OpenAI-compatible proxies via LLM__OPENAI_BASE_URL were previously ignored on the default path (#643, fixes #641)
• Internal N+1 query in dialectic agent tool execution — collapsed per-iteration DB lookups into a single fetch (#652)
• Dreamer threshold and time-guard semantics: count filter now includes only documents.level == 'explicit' (was inflating threshold via dreamer-created levels and creating a feedback loop); last_dream_at write relocated from enqueue to process so duplicate enqueues or failed runs no longer reset the 8-hour time guard (#573)
• Deriver: blank observations are filtered out before embedding (previously triggered noisy embedding calls and persisted empty rows) (#615)
• Surprisal module: filter format corrected from {"level": levels} to {"level": {"in": levels}} — the prior call silently returned 0 results and made the entire Surprisal phase of the Dream cycle a no-op (#581, fixes #559)
• Removed hardcoded stop_sequences override from Deriver ModelConfig (was clobbering operator-configured stop sequences) (#587)
• Embedding client: embed() now wraps single-string input in an array, restoring compatibility with OpenAI-compatible third-party providers that reject scalar input (#586)
• Docker Compose: deriver service startup gated on the API service healthcheck — prevents races where the deriver starts before the API has run migrations (#689)
• Docker image: HEALTHCHECK directive removed from the shared base image; service-level health checks now belong in each service’s own configuration (#530)
• Removed strict parameter validation for thinking params on Anthropic and OpenAI transports — was rejecting valid per-transport configs (#686)
• Stream-final retries pin to the AttemptPlan that succeeded rather than re-running provider selection through the outer current_attempt ContextVar (#459)
• Gemini cached_content reuse keys now include system_instruction and tool_config so cache hits don’t cross configurations (#459)
• CrewAI example updated for the latest CrewAI protocol (#631)

​

Removed

• src/utils/clients.py deleted; its responsibilities are split across src/llm/registry.py, src/llm/credentials.py, and the backend-specific modules (#459)
• HEALTHCHECK directive from the shared Docker image (#530)

​

Changed

• Tightened transaction scopes across search, agent tools, queue manager, and webhook delivery to minimize DB connection hold time during external operations (#525)
• Search operations refactored to two-phase pattern — external work (embeddings, LLM calls) completes before opening a transaction (#525)
• Agent tool executor performs external operations before acquiring DB sessions (#525)
• Queue manager transaction scope reduced to only the critical section (#525)
• Webhook delivery no longer holds a DB session parameter (#525)

​

Fixed

• Session leakage in non-session-scoped dialectic chat calls (#526)

​

Added

• Health check endpoint (/health) for container orchestration and load balancer probes (#510)

​

Fixed

• explicit rollback on all transactions to force connection closed

​

Added

• JSONB metadata validation enforces 100 key limit and max depth of 5 (#419)

​

Changed

• Schemas refactored from single schemas.py into schemas/api.py, schemas/configuration.py, and schemas/internal.py with backwards-compatible re-exports (#419)

​

Fixed

• Missing deleted_at filter on RepresentationManager._query_documents_recent() and ._query_documents_most_derived() allowed soft-deleted documents to leak into the deriver’s working representation (#456)
• CleanupStaleItemsCompletedEvent emitted spuriously when no queue item was actually deleted (#454)
• Empty JSON file uploads caused unhandled errors; now returns normalized error responses (#434)
• Memory leak: _observation_locks switched to WeakValueDictionary to prevent unbounded growth (#419)
• SQL injection in dependencies.py: parameterized set_config calls to prevent injection via request context (#419)
• NUL byte crashes: string inputs (message content, queries, peer cards) now stripped at schema level (#419)
• Filter recursion depth capped at 5 to prevent stack overflow (#419)
• Dedup-skipped observations now correctly reflected in created counts (#477)
• External vector store support for message search — routes queries through configured external vector store with oversampling and deduplication to handle chunked embeddings (#479)
• Dialectic agent no longer holds a DB connection during LLM calls — embeddings are pre-computed before tool execution, DB sessions isolated in extract_preferences, query_documents no longer accepts a DB session parameter (#477)

​

Added

• Consolidated session context into a single DB session with 40/60 token budget allocation between summary and messages
• Observation validation via ObservationInput Pydantic schema with partial-success support and batch embedding with per-observation fallback
• Peer card hard cap of 40 facts with case-insensitive deduplication and whitespace normalization
• Safe integer coercion (_safe_int) for all LLM tool inputs to handle non-integer values like "Infinity"
• Embedding pre-computation and reuse across multiple search calls in dialectic and representation flows
• Peer existence validation in dialectic chat endpoints — raises ResourceNotFoundException instead of silently failing
• Logging filter to suppress noisy GET /metrics access logs
• Oolong long-context aggregation benchmark (synth and real variants, 1K–4M token context windows)
• MolecularBench fact quality evaluation (ambiguity, decontextuality, minimality scoring)
• CoverageBench information recall evaluation (gold fact extraction, coverage matching, QA verification)
• LoCoMo summary-as-context baseline evaluation
• Webhook delivery tests, dependency lifecycle tests, queue cleanup tests, summarizer fallback tests
• Parallel test execution via pytest-xdist with worker-specific databases
• test_reasoning_levels.py script for LOCOM dataset testing across reasoning levels

​

Changed

• Workspace deletion is now async — returns 202 Accepted, validates no active sessions (409 Conflict), cascade-deletes in background
• Redis caching layer now stores plain-dict instead of ORM objects, with v2-prefixed keys, storage, resilient safe_cache_set/safe_cache_delete helpers, and deferred post-commit cache invalidation
• All get_or_create_* CRUD operations now use savepoints (db.begin_nested()) instead of commit/rollback for race condition prevention
• Reconciler vector sync uses direct ORM mutation instead of batch parameterized UPDATE statements
• Summarizer enforces hard word limit in prompt and creates fallback text for empty summaries with summary_tokens = 0
• Blocked Gemini responses (SAFETY, RECITATION, PROHIBITED_CONTENT, BLOCKLIST) now raise LLMError to trigger retry/backup-provider logic
• Gemini client explicitly sets max_output_tokens from max_tokens parameter
• All deriver and metrics collector logging replaced with structured logging.getLogger(__name__) calls
• Dreamer specialist prompts updated to enforce durable-facts-only peer cards with max 40 entries and deduplication
• GetOrCreateResult changed from NamedTuple to dataclass with async post_commit() method
• FastAPI upgraded from 0.111.0 to 0.131.0; added pyarrow dependency
• Queue status filtering to only show user-facing tasks (representation, summary, dream); excludes internal infrastructure tasks

​

Fixed

• JWT timestamp bug — JWTParams.t was evaluated once at class definition time instead of per-instance
• Session cache invalidation on deletion was missing
• get_peer_card() now properly propagates ResourceNotFoundException instead of swallowing it
• set_peer_card() ensures peer exists via get_or_create_peers() before updating
• Backup provider failover with proper tool input type safety
• Removed setup_admin_jwt() from server startup
• Sentry coroutine detection switched from asyncio.iscoroutinefunction to inspect.iscoroutinefunction

​

Removed

• explicit.py and obex.py benchmarks replaced by coverage.py and molecular.py
• Claude Code review automation workflow (.github/workflows/claude.yml)
• Coverage reporting from default pytest configuration

​

Added

• Documentation for reasoning_level and Claude Code plugin

​

Changed

• Gave dreaming sub-agents better prompting around peer card creation, tweaked overall prompts

​

Fixed

• Added message-search fallback for memory search tool, necessary in fresh sessions
• Made FLUSH_ENABLED a config value
• Removed N+1 query in search_messages

​

Fixed

• Token counting in Explicit Agent Loop
• Backwards compatibility of queue items

​

Added

• Agentic Dreamer for intelligent memory consolidation using LLM agents
• Agentic Dialectic for query answering using LLM agents with tool use
• Reasoning levels configuration for dialectic (minimal, low, medium, high, max)
• Prometheus token tracking for deriver and dialectic operations
• n8n integration
• Cloud Events for auditable telemetry
• External Vector Store support for turbopuffer and lancedb with reconciliation flow

​

Changed

• API route renaming for consistency
• Dreamer and dialectic now respect peer card configuration settings
• Observations renamed to Conclusions across API and SDKs
• Deriver to buffer representation tasks to normalize workloads
• Local Representation tasks to create singular QueueItems
• getContext endpoint to use search_query rather than force last_user_message

​

Fixed

• Dream scheduling bugs
• Summary creation when start_message_id > end_message_id
• Cashews upgrade to prevent NoScriptError
• Memory leak in accumulate_metric call

​

Removed

• Peer card configuration from message configuration; peer cards no longer created/updated in deriver process

​

Fixed

• Backwards compatibility for message_ids field in documents to handle legacy tuple format

​

Added

• Message level configurations
• CRUD operations for observations
• Comprehensive test cases for harness
• Peer level get_context
• Set Peer Card Method
• Manual dreaming trigger endpoint

​

Changed

• Configurations to support more flags for fine-grained control of the deriver, peer cards, summaries, etc.
• Working Representations to support more fine-grained parameters

​

Fixed

• File uploads to match MessageCreate structure
• Cache invalidation strategy

​

Added

• Redis caching to improve DB IO
• Backup LLM provider to avoid failures when a provider is down

​

Changed

• QueueItems to use standardized columns
• Improved Deduplication logic for Representation Tasks
• More finegrained metrics for representation, summary, and peer card tasks
• DB constraint to follow standard naming conventions

​

Fixed

• Langfuse tracing to have readable waterfalls
• Alembic Migrations to match models.py
• message_in_seq correctly included in webhook payload

​

Changed

• Alembic to always use a session pooler
• Statement timeout during alembic operations to 5 min

​

Added

• Alembic migration validation test suite

​

Fixed

• Alembic migrations to batch changes
• Batch message creation sequence number

​

Changed

• Logging infrastructure to remove noisy messages
• Sentry integration is centralized

​

Added

• Unified Representation class
• vllm client support
• Periodic queue cleanup logic
• WIP Dreaming Feature
• LongMemEval to Test Bench
• Prometheus Client for better Metrics
• Performance metrics instrumentation
• Error reporting to deriver
• Workspace Delete Method
• Multi-db option in test harness

​

Changed

• Working Representations are Queried on the fly rather than cached in metadata
• EmbeddingStore to RepresentationFactory
• Summary Response Model to use public_id of message for cutoff
• Semantic across codebase to reference resources based on observer and observed
• Prompts for Deriver & Dialectic to reference peer_id and add examples
• Get Context route returns peer card and representation in addition to messages and summaries
• Refactoring logger.info calls to logger.debug where applicable

​

Fixed

• Gemini client to use async methods

​

Changed

• Deriver Rollup Queue processes interleaved messages for more context

​

Fixed

• Dialectic Streaming to follow SSE conventions
• Sentry tracing in the deriver

​

Added

• Get peer cards endpoint (GET /v2/peers/{peer_id}/card) for retrieving targeted peer context information

​

Changed

• Replaced Mirascope dependency with small client implementation for better control
• Optimized deriver performance by using joins on messages table instead of storing token count in queue payload
• Database scope optimization for various operations
• Batch representation task processing for ~10x speed improvement in practice

​

Fixed

• Separated clean and claim work units in queue manager to prevent race conditions
• Skip locked ActiveQueueSession rows on delete operations
• Langfuse SDK integration updates for compatibility
• Added configurable maximum message size to prevent token overflow in deriver
• Various minor bugfixes

​

Fixed

• Added max message count to deriver in order to not overflow token limits

​

Added

• getSummaries endpoint to get all available summaries for a session directly
• Peer Card feature to improve context for deriver and dialectic

​

Changed

• Session Peer limit to be based on observers instead, renamed config value to SESSION_OBSERVERS_LIMIT
• Messages can take a custom timestamp for the created_at field, defaulting to the current time
• get_context endpoint returns detailed Summary object rather than just summary content
• Working representations use a FIFO queue structure to maintain facts rather than a full rewrite
• Optimized deriver enqueue by prefetching message sequence numbers (eliminates N+1 queries)

​

Fixed

• Deriver uses get_context internally to prevent context window limit errors
• Embedding store will truncate context when querying documents to prevent embedding token limit errors
• Queue manager to schedule work based on available works rather than total number of workers
• Queue manager to use atomic db transactions rather than long lived transaction for the worker lifecycle
• Timestamp formats unified to ISO 8601 across the codebase
• Internal get_context method’s cutoff value is exclusive now

​

Added

• Arbitrary filters now available on all search endpoints
• Search combines full-text and semantic using reciprocal rank fusion
• Webhook support (currently only supports queue_empty and test events, more to come)
• Small test harness and custom test format for evaluating Honcho output quality
• Added MCP server and documentation for it

​

Changed

• Search has 10 results by default, max 100 results
• Queue structure generalized to handle more event types
• Summarizer now exhaustive by default and tuned for performance

​

Fixed

• Resolve race condition for peers that leave a session while sending messages
• Added explicit rollback to solve integrity error in queue
• Re-introduced Sentry tracing to deriver
• Better integrity logic in get_or_create API methods

​

Fixed

• Summarizer module to ignore empty summaries and pass appropriate one to get_context
• Structured Outputs calls with OpenAI provider to pass strict=True to Pydantic Schema

​

Added

• Test harness for custom Honcho evaluations
• Better support for session and peer aware dialectic queries
• Langfuse settings
• Added recent history to dialectic prompt, dynamic based on new context window size setting

​

Fixed

• Summary queue logic
• Formatting of logs
• Filtering by session
• Peer targeting in queries

​

Changed

• Made query expansion in dialectic off by default
• Overhauled logging
• Refactor summarization for performance and code clarity
• Refactor queue payloads for clarity

​

Added

• File uploads
• Brand new “ROTE” deriver system
• Updated dialectic system
• Local working representations
• Better logging for deriver/dialectic
• Deriver Queue Status no longer has redundant data

​

Fixed

• Document insertion
• Session-scoped and peer-targeted dialectic queries work now
• Minor bugs

​

Removed

• Peer-level messages

​

Changed

• Dialectic chat endpoint takes a single query
• Rearranged configuration values (LLM, Deriver, Dialectic, History->Summary)

​

Fixed

• Groq API client to use the Async library

​

Fixed

• Migration/provision scripts did not have correct database connection arguments, causing timeouts

​

Fixed

• Bug that causes runtime error when Sentry flags are enabled

​

Fixed

• Database initialization was misconfigured and led to provision_db script failing: switch to consistent working configuration with transaction pooler

​

Added

• Ergonomic SDKs for Python and TypeScript (uses Stainless underneath)
• Deriver Queue Status endpoint
• Complex arbitrary filters on workspace/session/peer/message
• Message embedding table for full semantic search

​

Changed

• Overhauled documentation
• BasedPyright typing for entire project
• Resource filtering expanded to include logical operators

​

Fixed

• Various bugs
• Use new config arrangement everywhere
• Remove hardcoded responses

​

Added

• Ability to get a peer’s working representation
• Metadata to all data primitives (Workspaces, Peers, Sessions, Messages)
• Internal metadata to store Honcho’s state no longer exposed in API
• Batch message operations and enhanced message querying with token and message count limits
• Search and summary functionalities scoped by workspace, peer, and session
• Session context retrieval with summaries and token allocatio
• HNSW Index for Documents Table
• Centralized Configuration via Environment Variables or config.toml file

​

Changed

• New architecture centered around the concept of a “peer” replaces the former “app”/“user”/“session” paradigm
• Workspaces replace “apps” as top-level namespace
• Peers replace “users”
• Sessions no longer nested beneath peers and no longer limited to a single user-assistant model. A session exists independently of any one peer and peers can be added to and removed from sessions.
• Dialectic API is now part of the Peer, not the Session
• Dialectic API now allows queries to be scoped to a session or “targeted” to a fellow peer
• Database schema migrated to adopt workspace/peer/session naming and structure
• Authentication and JWT scopes updated to workspace/peer/session hierarchy
• Queue processing now works on ‘work units’ instead of sessions
• Message token counting updated with tiktoken integration and fallback heuristic
• Queue and message processing updated to handle sender/target and task types for multi-peer scenarios

​

Fixed

• Improved error handling and validation for batch message operations and metadata
• Database Sessions to be more atomic to reduce idle in transaction time

​

Removed

• Metamessages removed in favor of metadata
• Collections and Documents no longer exposed in the API, solely internal
• Obsolete tests for apps, users, collections, documents, and metamessages

​

Added

• Normalize resources to remove joins and increase query performance
• Query tracing for debugging

​

Changed

• /list endpoints to not require a request body
• metamessage_type to label with backwards compatibility
• Database Provisioning to rely on alembic
• Database Session Manager to explicitly rollback transactions before closing the connection

​

Fixed

• Alembic Migrations to include initial database migrations
• Sentry Middleware to not report Honcho Exceptions

​

Added

• JWT based API authentication
• Configurable logging
• Consolidated LLM Inference via ModelClient class
• Dynamic logging configurable via environment variables

​

Changed

• Deriver & Dialectic API to use Hybrid Memory Architecture
• Metamessages are not strictly tied to a message
• Database provisioning is a separate script instead of happening on startup
• Consolidated session/chat and session/chat/stream endpoints