Migrating from Parallel
Point your existing Parallel client at Brime — no code changes.
Brime exposes Parallel's wire format under /parallel/v1/*. Two endpoints (/search, /extract) supported in wave 1; /tasks, /findall, /monitor, /chat are on the roadmap.
One-line migration
from parallel import Parallel
client = Parallel(
api_key="sk-brime-...",
base_url="https://api.brime.dev/parallel", # ← only change
)
result = client.search.create(
objective="latest 2025 Federal Reserve rate decisions",
search_queries=["ECB rate forecast 2026", "EU inflation outlook"],
)import Parallel from "parallel-web";
const client = new Parallel({
apiKey: "sk-brime-...",
baseURL: "https://api.brime.dev/parallel",
});
const result = await client.search.create({
objective: "Tesla Cybertruck recall 2025",
search_queries: ["ECB rate forecast 2026"],
});What's compatible
| Parallel endpoint | Brime path | Status |
|---|---|---|
/search | /parallel/v1/search | ✅ Wave 1 |
/extract | /parallel/v1/extract | ✅ Wave 1 |
/tasks (deep async research) | — | ❌ Wave 2 |
/findall | — | ❌ Wave 2 |
/monitor (scheduled + webhooks) | — | ❌ Wave 2 |
/chat | — | ❌ Wave 2 |
Field mapping
| Parallel field | Brime behaviour |
|---|---|
objective + search_queries[] | If objective is set, used as the Brime query; otherwise search_queries.join(" OR ") |
mode: basic / advanced | Maps to Brime depth=basic / depth=advanced |
client_model | Silently ignored |
advanced_settings | Silently ignored |
max_chars_total | Silently ignored |
excerpts[] per result | Returned as a single 1,000-char slice (not multi-chunk yet) |
full_content (extract) | Brime markdown |
Authentication
Both x-api-key: <key> (Parallel default) and Authorization: Bearer <key> work.
ID prefixes
Brime synthesises IDs that match Parallel's prefix convention:
- Search:
search_<request_id> - Extract:
extract_<request_id> - Session:
session_<uuid>(echoed when you supply one)
Pricing
Brime credits. See Pricing. Parallel's processor tiers (lite/base/core/pro/ultra) are not emulated; Brime's depth presets are simpler.
When to use the Parallel adapter vs the native brime SDK
- Use the Parallel adapter to keep an existing Parallel SDK call site working.
- Use the native
brimeSDK when starting fresh — covers all three native endpoints (search, extract, research) under one surface.